2020-02-01 16:55:41.341: INFO @api : Creating dataset 2020-02-01 16:55:42.368: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/utils/profiler.py:12: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead. 2020-02-01 16:55:42.700: INFO @api : RecognitionDatasetWrapper created 2020-02-01 16:55:42.700: INFO @api : Creating a model 2020-02-01 16:55:42.720: INFO @model : Creating TF model on 1 GPU devices 2020-02-01 16:55:42.720: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:415: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead. 2020-02-01 16:55:43.174: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:140: The name tf.variable_scope is deprecated. Please use tf.compat.v1.variable_scope instead. 2020-02-01 16:55:43.175: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:140: The name tf.get_variable_scope is deprecated. Please use tf.compat.v1.get_variable_scope instead. 2020-02-01 16:55:43.175: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:140: The name tf.AUTO_REUSE is deprecated. Please use tf.compat.v1.AUTO_REUSE instead. 2020-02-01 16:55:43.245: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.268: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.355: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.377: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.461: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.484: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.546: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.607: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.629: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.713: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.735: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.824: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.848: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.911: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.974: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:43.997: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.081: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.105: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.280: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.303: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.367: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.431: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.454: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.539: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.561: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.647: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.670: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.732: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.793: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.815: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.841: WARNING @deprecation : From /home/ala/faces/models/recognition.py:81: flatten (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.flatten instead. 2020-02-01 16:55:44.851: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AttributeError: module 'gast' has no attribute 'Num' 2020-02-01 16:55:44.855: WARNING @deprecation : From /home/ala/faces/models/recognition.py:82: dropout (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dropout instead. 2020-02-01 16:55:44.868: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.879: WARNING @deprecation : From /home/ala/faces/models/recognition.py:84: dense (from tensorflow.python.layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dense instead. 2020-02-01 16:55:44.880: WARNING @deprecation : From /usr/lib/python3.7/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version. Instructions for updating: Call initializer instance with the dtype argument instead of passing it to the constructor 2020-02-01 16:55:44.960: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:44.974: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:45.065: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:45.148: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:45.163: WARNING @deprecation : From /home/ala/faces/models/recognition.py:45: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version. Instructions for updating: Use tf.where in 2.0, which has the same broadcast rule as np.where 2020-02-01 16:55:45.247: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:45.329: WARNING @ag_logging : Entity > could not be transformed and will be executed as-is. Please report this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: converting >: AssertionError: Bad argument number for Name: 3, expecting 4 2020-02-01 16:55:46.738: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:180: The name tf.train.Saver is deprecated. Please use tf.compat.v1.train.Saver instead. 2020-02-01 16:55:47.316: INFO @model : Number of parameters: 1010079 2020-02-01 16:55:47.316: INFO @api : RecognitionModel created 2020-02-01 16:55:47.317: INFO @api : Creating hooks 2020-02-01 16:55:47.326: INFO @api : ComputeStats created 2020-02-01 16:55:47.327: INFO @api : LogVariables created 2020-02-01 16:55:47.327: INFO @api : LogProfile created 2020-02-01 16:55:47.328: INFO @api : LogDir created 2020-02-01 16:55:47.336: WARNING @abstract_hook: Argument `n_epochs` was not recognized by `ComputeRecognitionMetrics`. Recognized arguments are `['output_dir', 'skip_epochs', 'streams', 'kwargs']`. 2020-02-01 16:55:47.336: INFO @api : ComputeRecognitionMetrics created 2020-02-01 16:55:47.337: INFO @api : EvaluateConfidence created 2020-02-01 16:55:47.337: INFO @api : EvaluateGenderAgeModel created 2020-02-01 16:55:47.337: INFO @api : EvaluateGenderAgeModel created 2020-02-01 16:55:47.338: INFO @api : SaveAgePredictions created 2020-02-01 16:55:47.338: INFO @api : SaveGenderPredictions created 2020-02-01 16:55:47.338: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/hooks/write_tensorboard.py:78: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead. 2020-02-01 16:55:47.338: INFO @api : WriteTensorBoard created 2020-02-01 16:55:47.339: INFO @api : DecayLR created 2020-02-01 16:55:47.339: INFO @api : SaveBest created 2020-02-01 16:55:47.340: INFO @api : StopAfter created 2020-02-01 16:55:47.341: INFO @api : ShowProgress created 2020-02-01 16:55:47.341: WARNING @api : TrainingTrace hook added between hooks. Add it to your config.yaml to suppress this warning. 2020-02-01 16:55:47.341: INFO @api : Creating main loop 2020-02-01 16:55:47.341: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 16:55:47.342: INFO @main_loop : Training epoch 1 2020-02-01 16:55:54.419: WARNING @main_loop : Some sources provided by the stream do not match model placeholders. Set `main_loop.on_unused_sources` to `ignore` in order to suppress this warning. Extra sources: ['box_points', 'orig_box', 'filenames'] 2020-02-01 16:58:08.269: INFO @log_variables: train loss nanmean: 3.225131 2020-02-01 16:58:08.269: INFO @log_variables: train age_loss mean: 27.189421 2020-02-01 16:58:08.269: INFO @log_variables: train gender_loss mean: 0.656437 2020-02-01 16:58:08.269: INFO @log_variables: train age_mae mean: 27.685915 2020-02-01 16:58:08.269: INFO @log_variables: train gender_accuracy mean: 0.653592 2020-02-01 16:58:08.269: INFO @log_variables: train gender_confidence/loss nanmean: 0.108531 2020-02-01 16:58:08.270: INFO @log_variables: train gender_confidence/accuracy mean: 0.542544 2020-02-01 16:58:08.270: INFO @log_variables: train age_confidence/loss mean: 0.048073 2020-02-01 16:58:08.270: INFO @log_variables: train age_confidence/accuracy mean: 0.728181 2020-02-01 16:58:08.270: INFO @log_variables: valid loss nanmean: 1.925657 2020-02-01 16:58:08.270: INFO @log_variables: valid age_loss mean: 14.113873 2020-02-01 16:58:08.270: INFO @log_variables: valid gender_loss mean: 0.556029 2020-02-01 16:58:08.270: INFO @log_variables: valid age_mae mean: 14.604501 2020-02-01 16:58:08.270: INFO @log_variables: valid gender_accuracy mean: 0.719882 2020-02-01 16:58:08.270: INFO @log_variables: valid gender_confidence/loss nanmean: 0.082364 2020-02-01 16:58:08.270: INFO @log_variables: valid gender_confidence/accuracy mean: 0.632576 2020-02-01 16:58:08.270: INFO @log_variables: valid age_confidence/loss mean: 0.054733 2020-02-01 16:58:08.270: INFO @log_variables: valid age_confidence/accuracy mean: 0.726685 2020-02-01 16:58:08.270: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 16:58:08.277: INFO @metrics_hook: train age_mae: 27.686 +-0.116 (110592) 2020-02-01 16:58:08.285: INFO @metrics_hook: train gender_accuracy: 0.654 +-0.003 (110592) 2020-02-01 16:58:11.187: INFO @metrics_hook: valid age_mae: 14.605 +-0.202 (17639) 2020-02-01 16:58:11.188: INFO @metrics_hook: valid gender_accuracy: 0.720 +-0.007 (17639) 2020-02-01 16:58:12.849: INFO @decay_lr : LR updated to `9.95e-05` 2020-02-01 16:58:12.849: WARNING @deprecation_wrapper: From /home/ala/emloop-tensorflow/emloop_tensorflow/model.py:301: The name tf.train.write_graph is deprecated. Please use tf.io.write_graph instead. 2020-02-01 16:58:13.259: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 16:58:13.260: INFO @log_profile : T train: 123.571686 2020-02-01 16:58:13.260: INFO @log_profile : T valid: 5.841029 2020-02-01 16:58:13.261: INFO @log_profile : T read data: 8.992821 2020-02-01 16:58:13.261: INFO @log_profile : T hooks: 7.423442 2020-02-01 16:58:13.261: INFO @main_loop : Epoch 1 done 2020-02-01 16:58:13.261: INFO @main_loop : Training epoch 2 2020-02-01 17:00:34.050: INFO @log_variables: train loss nanmean: 2.028445 2020-02-01 17:00:34.050: INFO @log_variables: train age_loss mean: 14.841405 2020-02-01 17:00:34.050: INFO @log_variables: train gender_loss mean: 0.560501 2020-02-01 17:00:34.050: INFO @log_variables: train age_mae mean: 15.334344 2020-02-01 17:00:34.050: INFO @log_variables: train gender_accuracy mean: 0.716187 2020-02-01 17:00:34.050: INFO @log_variables: train gender_confidence/loss nanmean: 0.106935 2020-02-01 17:00:34.050: INFO @log_variables: train gender_confidence/accuracy mean: 0.586045 2020-02-01 17:00:34.050: INFO @log_variables: train age_confidence/loss mean: 0.062745 2020-02-01 17:00:34.050: INFO @log_variables: train age_confidence/accuracy mean: 0.652584 2020-02-01 17:00:34.050: INFO @log_variables: valid loss nanmean: 1.458131 2020-02-01 17:00:34.050: INFO @log_variables: valid age_loss mean: 9.974669 2020-02-01 17:00:34.050: INFO @log_variables: valid gender_loss mean: 0.459632 2020-02-01 17:00:34.050: INFO @log_variables: valid age_mae mean: 10.463450 2020-02-01 17:00:34.050: INFO @log_variables: valid gender_accuracy mean: 0.768921 2020-02-01 17:00:34.050: INFO @log_variables: valid gender_confidence/loss nanmean: 0.076631 2020-02-01 17:00:34.050: INFO @log_variables: valid gender_confidence/accuracy mean: 0.679404 2020-02-01 17:00:34.050: INFO @log_variables: valid age_confidence/loss mean: 0.056865 2020-02-01 17:00:34.050: INFO @log_variables: valid age_confidence/accuracy mean: 0.681955 2020-02-01 17:00:34.051: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:00:34.058: INFO @metrics_hook: train age_mae: 15.334 +-0.077 (110372) 2020-02-01 17:00:34.065: INFO @metrics_hook: train gender_accuracy: 0.716 +-0.003 (110372) 2020-02-01 17:00:36.800: INFO @metrics_hook: valid age_mae: 10.463 +-0.132 (17639) 2020-02-01 17:00:36.801: INFO @metrics_hook: valid gender_accuracy: 0.769 +-0.006 (17639) 2020-02-01 17:00:38.265: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:00:38.266: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.62 +- 0.29 2020-02-01 17:00:38.266: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.56 +- 0.27 2020-02-01 17:00:38.266: INFO @evaluate_confidence: Average confidence of all samples 0.60 +- 0.28 2020-02-01 17:00:38.415: INFO @evaluate_confidence: Previous accuracy would be: 71.62 2020-02-01 17:00:38.416: INFO @evaluate_confidence: Possible optimal thresholds are: [0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62] 2020-02-01 17:00:38.430: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [74.33, 74.42, 74.52, 74.66, 74.79, 74.95, 75.11] 2020-02-01 17:00:38.430: INFO @evaluate_confidence: Dropped ratios are: [42.57, 43.65, 44.72, 45.74, 46.87, 47.99, 49.07] 2020-02-01 17:00:38.480: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:00:38.480: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.32 +- 0.27 2020-02-01 17:00:38.480: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.26 2020-02-01 17:00:38.480: INFO @evaluate_confidence: Average confidence of all samples 0.32 +- 0.26 2020-02-01 17:00:38.607: INFO @evaluate_confidence: Previous accuracy would be: 22.66 2020-02-01 17:00:38.607: INFO @evaluate_confidence: Possible optimal thresholds are: [] 2020-02-01 17:00:38.607: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [] 2020-02-01 17:00:38.607: INFO @evaluate_confidence: Dropped ratios are: [] 2020-02-01 17:00:38.614: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:00:38.614: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.70 +- 0.17 2020-02-01 17:00:38.615: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.61 +- 0.10 2020-02-01 17:00:38.615: INFO @evaluate_confidence: Average confidence of all samples 0.68 +- 0.16 2020-02-01 17:00:38.922: INFO @evaluate_confidence: Previous accuracy would be: 76.89 2020-02-01 17:00:38.922: INFO @evaluate_confidence: Possible optimal thresholds are: [0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69] 2020-02-01 17:00:38.924: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [82.15, 82.97, 83.77, 84.79, 85.53, 86.3, 86.98, 87.96, 88.8] 2020-02-01 17:00:38.924: INFO @evaluate_confidence: Dropped ratios are: [41.93, 43.98, 45.85, 47.75, 49.69, 51.52, 53.15, 54.88, 56.43] 2020-02-01 17:00:38.931: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:00:38.932: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.30 +- 0.05 2020-02-01 17:00:38.932: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.30 +- 0.05 2020-02-01 17:00:38.932: INFO @evaluate_confidence: Average confidence of all samples 0.30 +- 0.05 2020-02-01 17:00:39.053: INFO @evaluate_confidence: Previous accuracy would be: 31.92 2020-02-01 17:00:39.053: INFO @evaluate_confidence: Possible optimal thresholds are: [] 2020-02-01 17:00:39.053: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [] 2020-02-01 17:00:39.053: INFO @evaluate_confidence: Dropped ratios are: [] 2020-02-01 17:00:39.100: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:00:39.756: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:00:39.839: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:00:40.279: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:00:40.486: INFO @decay_lr : LR updated to `9.90025e-05` 2020-02-01 17:00:40.763: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:00:40.764: INFO @log_profile : T train: 130.584368 2020-02-01 17:00:40.764: INFO @log_profile : T valid: 6.690382 2020-02-01 17:00:40.764: INFO @log_profile : T read data: 2.864753 2020-02-01 17:00:40.764: INFO @log_profile : T hooks: 7.287323 2020-02-01 17:00:40.764: INFO @main_loop : Epoch 2 done 2020-02-01 17:00:40.764: INFO @main_loop : Training epoch 3 2020-02-01 17:03:00.880: INFO @log_variables: train loss nanmean: 1.723891 2020-02-01 17:03:00.880: INFO @log_variables: train age_loss mean: 12.287316 2020-02-01 17:03:00.880: INFO @log_variables: train gender_loss mean: 0.501818 2020-02-01 17:03:00.880: INFO @log_variables: train age_mae mean: 12.778553 2020-02-01 17:03:00.880: INFO @log_variables: train gender_accuracy mean: 0.748436 2020-02-01 17:03:00.880: INFO @log_variables: train gender_confidence/loss nanmean: 0.091824 2020-02-01 17:03:00.880: INFO @log_variables: train gender_confidence/accuracy mean: 0.633644 2020-02-01 17:03:00.880: INFO @log_variables: train age_confidence/loss mean: 0.058840 2020-02-01 17:03:00.880: INFO @log_variables: train age_confidence/accuracy mean: 0.640842 2020-02-01 17:03:00.880: INFO @log_variables: valid loss nanmean: 1.336777 2020-02-01 17:03:00.880: INFO @log_variables: valid age_loss mean: 8.994308 2020-02-01 17:03:00.880: INFO @log_variables: valid gender_loss mean: 0.422888 2020-02-01 17:03:00.880: INFO @log_variables: valid age_mae mean: 9.482151 2020-02-01 17:03:00.880: INFO @log_variables: valid gender_accuracy mean: 0.794886 2020-02-01 17:03:00.880: INFO @log_variables: valid gender_confidence/loss nanmean: 0.076094 2020-02-01 17:03:00.881: INFO @log_variables: valid gender_confidence/accuracy mean: 0.767674 2020-02-01 17:03:00.881: INFO @log_variables: valid age_confidence/loss mean: 0.058575 2020-02-01 17:03:00.881: INFO @log_variables: valid age_confidence/accuracy mean: 0.654289 2020-02-01 17:03:00.881: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:03:00.888: INFO @metrics_hook: train age_mae: 12.779 +-0.062 (110592) 2020-02-01 17:03:00.895: INFO @metrics_hook: train gender_accuracy: 0.748 +-0.003 (110592) 2020-02-01 17:03:03.578: INFO @metrics_hook: valid age_mae: 9.482 +-0.116 (17639) 2020-02-01 17:03:03.579: INFO @metrics_hook: valid gender_accuracy: 0.795 +-0.006 (17639) 2020-02-01 17:03:05.408: INFO @decay_lr : LR updated to `9.850749e-05` 2020-02-01 17:03:05.686: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:03:05.688: INFO @log_profile : T train: 130.728824 2020-02-01 17:03:05.688: INFO @log_profile : T valid: 6.869145 2020-02-01 17:03:05.688: INFO @log_profile : T read data: 1.846406 2020-02-01 17:03:05.688: INFO @log_profile : T hooks: 5.402495 2020-02-01 17:03:05.689: INFO @main_loop : Epoch 3 done 2020-02-01 17:03:05.689: INFO @main_loop : Training epoch 4 2020-02-01 17:05:26.420: INFO @log_variables: train loss nanmean: 1.575548 2020-02-01 17:05:26.420: INFO @log_variables: train age_loss mean: 11.152771 2020-02-01 17:05:26.420: INFO @log_variables: train gender_loss mean: 0.461805 2020-02-01 17:05:26.420: INFO @log_variables: train age_mae mean: 11.643112 2020-02-01 17:05:26.420: INFO @log_variables: train gender_accuracy mean: 0.771636 2020-02-01 17:05:26.420: INFO @log_variables: train gender_confidence/loss nanmean: 0.084327 2020-02-01 17:05:26.420: INFO @log_variables: train gender_confidence/accuracy mean: 0.662786 2020-02-01 17:05:26.420: INFO @log_variables: train age_confidence/loss mean: 0.057510 2020-02-01 17:05:26.421: INFO @log_variables: train age_confidence/accuracy mean: 0.644004 2020-02-01 17:05:26.421: INFO @log_variables: valid loss nanmean: 1.299363 2020-02-01 17:05:26.421: INFO @log_variables: valid age_loss mean: 8.642890 2020-02-01 17:05:26.421: INFO @log_variables: valid gender_loss mean: 0.421541 2020-02-01 17:05:26.421: INFO @log_variables: valid age_mae mean: 9.129657 2020-02-01 17:05:26.421: INFO @log_variables: valid gender_accuracy mean: 0.801236 2020-02-01 17:05:26.421: INFO @log_variables: valid gender_confidence/loss nanmean: 0.072429 2020-02-01 17:05:26.421: INFO @log_variables: valid gender_confidence/accuracy mean: 0.775044 2020-02-01 17:05:26.421: INFO @log_variables: valid age_confidence/loss mean: 0.057997 2020-02-01 17:05:26.421: INFO @log_variables: valid age_confidence/accuracy mean: 0.647826 2020-02-01 17:05:26.421: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:05:26.429: INFO @metrics_hook: train age_mae: 11.643 +-0.057 (110372) 2020-02-01 17:05:26.436: INFO @metrics_hook: train gender_accuracy: 0.772 +-0.003 (110372) 2020-02-01 17:05:29.255: INFO @metrics_hook: valid age_mae: 9.130 +-0.113 (17639) 2020-02-01 17:05:29.256: INFO @metrics_hook: valid gender_accuracy: 0.801 +-0.006 (17639) 2020-02-01 17:05:30.727: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:05:30.728: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.65 +- 0.22 2020-02-01 17:05:30.728: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.55 +- 0.19 2020-02-01 17:05:30.728: INFO @evaluate_confidence: Average confidence of all samples 0.63 +- 0.21 2020-02-01 17:05:30.870: INFO @evaluate_confidence: Previous accuracy would be: 77.16 2020-02-01 17:05:30.870: INFO @evaluate_confidence: Possible optimal thresholds are: [0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64] 2020-02-01 17:05:30.889: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [82.28, 82.53, 82.84, 83.09, 83.38, 83.74, 84.0, 84.37, 84.75, 85.04] 2020-02-01 17:05:30.890: INFO @evaluate_confidence: Dropped ratios are: [38.0, 39.77, 41.47, 43.09, 44.7, 46.3, 47.92, 49.55, 51.13, 52.68] 2020-02-01 17:05:30.939: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:05:30.939: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.38 +- 0.20 2020-02-01 17:05:30.939: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.18 2020-02-01 17:05:30.940: INFO @evaluate_confidence: Average confidence of all samples 0.37 +- 0.18 2020-02-01 17:05:31.082: INFO @evaluate_confidence: Previous accuracy would be: 28.73 2020-02-01 17:05:31.082: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37] 2020-02-01 17:05:31.087: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [30.64, 30.81] 2020-02-01 17:05:31.087: INFO @evaluate_confidence: Dropped ratios are: [54.41, 56.71] 2020-02-01 17:05:31.095: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:05:31.095: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.70 +- 0.16 2020-02-01 17:05:31.095: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.57 +- 0.09 2020-02-01 17:05:31.095: INFO @evaluate_confidence: Average confidence of all samples 0.67 +- 0.16 2020-02-01 17:05:31.218: INFO @evaluate_confidence: Previous accuracy would be: 80.12 2020-02-01 17:05:31.218: INFO @evaluate_confidence: Possible optimal thresholds are: [0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69] 2020-02-01 17:05:31.221: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [88.0, 88.52, 89.2, 89.87, 90.55, 91.24, 91.79, 92.31, 92.91, 93.37, 93.79, 94.24, 94.67] 2020-02-01 17:05:31.221: INFO @evaluate_confidence: Dropped ratios are: [34.17, 36.72, 39.11, 41.61, 44.01, 46.41, 48.75, 50.63, 52.67, 54.45, 56.34, 58.0, 59.81] 2020-02-01 17:05:31.229: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:05:31.229: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.36 +- 0.07 2020-02-01 17:05:31.229: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.06 2020-02-01 17:05:31.229: INFO @evaluate_confidence: Average confidence of all samples 0.34 +- 0.06 2020-02-01 17:05:31.358: INFO @evaluate_confidence: Previous accuracy would be: 35.26 2020-02-01 17:05:31.358: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36] 2020-02-01 17:05:31.359: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [42.86, 44.68, 45.81] 2020-02-01 17:05:31.359: INFO @evaluate_confidence: Dropped ratios are: [55.38, 64.15, 71.68] 2020-02-01 17:05:31.409: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:05:32.091: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:05:32.177: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:05:32.601: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:05:32.819: INFO @decay_lr : LR updated to `9.801495e-05` 2020-02-01 17:05:33.428: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:05:33.429: INFO @log_profile : T train: 130.439213 2020-02-01 17:05:33.429: INFO @log_profile : T valid: 6.829466 2020-02-01 17:05:33.429: INFO @log_profile : T read data: 2.782553 2020-02-01 17:05:33.430: INFO @log_profile : T hooks: 7.614346 2020-02-01 17:05:33.430: INFO @main_loop : Epoch 4 done 2020-02-01 17:05:33.430: INFO @main_loop : Training epoch 5 2020-02-01 17:07:54.569: INFO @log_variables: train loss nanmean: 1.486988 2020-02-01 17:07:54.569: INFO @log_variables: train age_loss mean: 10.580016 2020-02-01 17:07:54.569: INFO @log_variables: train gender_loss mean: 0.429862 2020-02-01 17:07:54.569: INFO @log_variables: train age_mae mean: 11.069631 2020-02-01 17:07:54.569: INFO @log_variables: train gender_accuracy mean: 0.790400 2020-02-01 17:07:54.569: INFO @log_variables: train gender_confidence/loss nanmean: 0.078262 2020-02-01 17:07:54.569: INFO @log_variables: train gender_confidence/accuracy mean: 0.698411 2020-02-01 17:07:54.569: INFO @log_variables: train age_confidence/loss mean: 0.056124 2020-02-01 17:07:54.570: INFO @log_variables: train age_confidence/accuracy mean: 0.653073 2020-02-01 17:07:54.570: INFO @log_variables: valid loss nanmean: 1.303015 2020-02-01 17:07:54.570: INFO @log_variables: valid age_loss mean: 9.066397 2020-02-01 17:07:54.570: INFO @log_variables: valid gender_loss mean: 0.388652 2020-02-01 17:07:54.570: INFO @log_variables: valid age_mae mean: 9.553945 2020-02-01 17:07:54.570: INFO @log_variables: valid gender_accuracy mean: 0.814899 2020-02-01 17:07:54.570: INFO @log_variables: valid gender_confidence/loss nanmean: 0.069679 2020-02-01 17:07:54.570: INFO @log_variables: valid gender_confidence/accuracy mean: 0.753501 2020-02-01 17:07:54.570: INFO @log_variables: valid age_confidence/loss mean: 0.055799 2020-02-01 17:07:54.570: INFO @log_variables: valid age_confidence/accuracy mean: 0.657237 2020-02-01 17:07:54.570: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:07:54.577: INFO @metrics_hook: train age_mae: 11.070 +-0.055 (110372) 2020-02-01 17:07:54.585: INFO @metrics_hook: train gender_accuracy: 0.790 +-0.002 (110372) 2020-02-01 17:07:57.273: INFO @metrics_hook: valid age_mae: 9.554 +-0.113 (17639) 2020-02-01 17:07:57.274: INFO @metrics_hook: valid gender_accuracy: 0.815 +-0.006 (17639) 2020-02-01 17:07:58.826: INFO @decay_lr : LR updated to `9.752488e-05` 2020-02-01 17:07:59.109: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:07:59.111: INFO @log_profile : T train: 130.832583 2020-02-01 17:07:59.111: INFO @log_profile : T valid: 6.811143 2020-02-01 17:07:59.111: INFO @log_profile : T read data: 2.804778 2020-02-01 17:07:59.111: INFO @log_profile : T hooks: 5.155475 2020-02-01 17:07:59.111: INFO @main_loop : Epoch 5 done 2020-02-01 17:07:59.111: INFO @main_loop : Training epoch 6 2020-02-01 17:10:15.051: INFO @log_variables: train loss nanmean: 1.439722 2020-02-01 17:10:15.051: INFO @log_variables: train age_loss mean: 10.311123 2020-02-01 17:10:15.051: INFO @log_variables: train gender_loss mean: 0.408969 2020-02-01 17:10:15.051: INFO @log_variables: train age_mae mean: 10.800171 2020-02-01 17:10:15.051: INFO @log_variables: train gender_accuracy mean: 0.803024 2020-02-01 17:10:15.052: INFO @log_variables: train gender_confidence/loss nanmean: 0.075323 2020-02-01 17:10:15.052: INFO @log_variables: train gender_confidence/accuracy mean: 0.715251 2020-02-01 17:10:15.052: INFO @log_variables: train age_confidence/loss mean: 0.055234 2020-02-01 17:10:15.052: INFO @log_variables: train age_confidence/accuracy mean: 0.658917 2020-02-01 17:10:15.052: INFO @log_variables: valid loss nanmean: 1.252601 2020-02-01 17:10:15.052: INFO @log_variables: valid age_loss mean: 8.756265 2020-02-01 17:10:15.052: INFO @log_variables: valid gender_loss mean: 0.364556 2020-02-01 17:10:15.052: INFO @log_variables: valid age_mae mean: 9.245237 2020-02-01 17:10:15.052: INFO @log_variables: valid gender_accuracy mean: 0.832700 2020-02-01 17:10:15.052: INFO @log_variables: valid gender_confidence/loss nanmean: 0.068152 2020-02-01 17:10:15.052: INFO @log_variables: valid gender_confidence/accuracy mean: 0.730314 2020-02-01 17:10:15.052: INFO @log_variables: valid age_confidence/loss mean: 0.057010 2020-02-01 17:10:15.052: INFO @log_variables: valid age_confidence/accuracy mean: 0.647939 2020-02-01 17:10:15.052: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:10:15.059: INFO @metrics_hook: train age_mae: 10.800 +-0.054 (110592) 2020-02-01 17:10:15.066: INFO @metrics_hook: train gender_accuracy: 0.803 +-0.002 (110592) 2020-02-01 17:10:17.779: INFO @metrics_hook: valid age_mae: 9.245 +-0.110 (17639) 2020-02-01 17:10:17.780: INFO @metrics_hook: valid gender_accuracy: 0.833 +-0.006 (17639) 2020-02-01 17:10:19.197: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:10:19.198: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.68 +- 0.21 2020-02-01 17:10:19.198: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.52 +- 0.15 2020-02-01 17:10:19.198: INFO @evaluate_confidence: Average confidence of all samples 0.65 +- 0.21 2020-02-01 17:10:19.340: INFO @evaluate_confidence: Previous accuracy would be: 80.30 2020-02-01 17:10:19.341: INFO @evaluate_confidence: Possible optimal thresholds are: [0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68] 2020-02-01 17:10:19.373: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [87.21, 87.64, 88.04, 88.44, 88.86, 89.25, 89.67, 90.07, 90.47, 90.82, 91.2, 91.53, 91.9, 92.21, 92.57, 92.88, 93.17] 2020-02-01 17:10:19.373: INFO @evaluate_confidence: Dropped ratios are: [32.28, 34.09, 35.81, 37.49, 39.16, 40.82, 42.51, 44.12, 45.66, 47.22, 48.66, 50.1, 51.52, 52.9, 54.26, 55.56, 56.84] 2020-02-01 17:10:19.424: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:10:19.424: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.41 +- 0.15 2020-02-01 17:10:19.424: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.13 2020-02-01 17:10:19.424: INFO @evaluate_confidence: Average confidence of all samples 0.39 +- 0.13 2020-02-01 17:10:19.563: INFO @evaluate_confidence: Previous accuracy would be: 31.71 2020-02-01 17:10:19.563: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4] 2020-02-01 17:10:19.572: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [36.6, 37.11, 37.6, 38.24] 2020-02-01 17:10:19.572: INFO @evaluate_confidence: Dropped ratios are: [53.49, 57.11, 60.45, 63.66] 2020-02-01 17:10:19.579: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:10:19.580: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.69 +- 0.21 2020-02-01 17:10:19.580: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.13 2020-02-01 17:10:19.580: INFO @evaluate_confidence: Average confidence of all samples 0.66 +- 0.21 2020-02-01 17:10:19.680: INFO @evaluate_confidence: Previous accuracy would be: 83.27 2020-02-01 17:10:19.680: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68] 2020-02-01 17:10:19.685: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [90.55, 90.82, 91.38, 91.93, 92.37, 92.72, 93.12, 93.48, 93.87, 94.26, 94.56, 94.88, 95.12, 95.3, 95.57, 95.74, 95.92, 96.1, 96.2, 96.39] 2020-02-01 17:10:19.685: INFO @evaluate_confidence: Dropped ratios are: [29.33, 31.03, 32.84, 34.49, 36.23, 37.69, 39.33, 40.84, 42.2, 43.58, 44.97, 46.2, 47.46, 48.59, 49.87, 50.93, 52.16, 53.21, 54.36, 55.58] 2020-02-01 17:10:19.693: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:10:19.693: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.38 +- 0.08 2020-02-01 17:10:19.693: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.07 2020-02-01 17:10:19.694: INFO @evaluate_confidence: Average confidence of all samples 0.37 +- 0.07 2020-02-01 17:10:19.815: INFO @evaluate_confidence: Previous accuracy would be: 34.53 2020-02-01 17:10:19.815: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37] 2020-02-01 17:10:19.816: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [40.49, 41.04] 2020-02-01 17:10:19.816: INFO @evaluate_confidence: Dropped ratios are: [52.75, 57.93] 2020-02-01 17:10:19.868: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:10:20.537: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:10:20.618: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:10:21.040: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:10:21.259: INFO @decay_lr : LR updated to `9.703725e-05` 2020-02-01 17:10:21.949: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:10:21.951: INFO @log_profile : T train: 127.665971 2020-02-01 17:10:21.951: INFO @log_profile : T valid: 5.662340 2020-02-01 17:10:21.951: INFO @log_profile : T read data: 1.927708 2020-02-01 17:10:21.951: INFO @log_profile : T hooks: 7.506894 2020-02-01 17:10:21.951: INFO @main_loop : Epoch 6 done 2020-02-01 17:10:21.951: INFO @main_loop : Training epoch 7 2020-02-01 17:12:33.192: INFO @log_variables: train loss nanmean: 1.388809 2020-02-01 17:12:33.193: INFO @log_variables: train age_loss mean: 9.952538 2020-02-01 17:12:33.193: INFO @log_variables: train gender_loss mean: 0.390525 2020-02-01 17:12:33.193: INFO @log_variables: train age_mae mean: 10.441083 2020-02-01 17:12:33.193: INFO @log_variables: train gender_accuracy mean: 0.814400 2020-02-01 17:12:33.193: INFO @log_variables: train gender_confidence/loss nanmean: 0.073598 2020-02-01 17:12:33.193: INFO @log_variables: train gender_confidence/accuracy mean: 0.724423 2020-02-01 17:12:33.193: INFO @log_variables: train age_confidence/loss mean: 0.055411 2020-02-01 17:12:33.193: INFO @log_variables: train age_confidence/accuracy mean: 0.659397 2020-02-01 17:12:33.193: INFO @log_variables: valid loss nanmean: 1.209788 2020-02-01 17:12:33.193: INFO @log_variables: valid age_loss mean: 8.239676 2020-02-01 17:12:33.193: INFO @log_variables: valid gender_loss mean: 0.366708 2020-02-01 17:12:33.193: INFO @log_variables: valid age_mae mean: 8.727170 2020-02-01 17:12:33.193: INFO @log_variables: valid gender_accuracy mean: 0.827655 2020-02-01 17:12:33.193: INFO @log_variables: valid gender_confidence/loss nanmean: 0.067494 2020-02-01 17:12:33.193: INFO @log_variables: valid gender_confidence/accuracy mean: 0.776518 2020-02-01 17:12:33.193: INFO @log_variables: valid age_confidence/loss mean: 0.059863 2020-02-01 17:12:33.193: INFO @log_variables: valid age_confidence/accuracy mean: 0.622201 2020-02-01 17:12:33.193: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:12:33.200: INFO @metrics_hook: train age_mae: 10.441 +-0.053 (110372) 2020-02-01 17:12:33.207: INFO @metrics_hook: train gender_accuracy: 0.814 +-0.002 (110372) 2020-02-01 17:12:35.924: INFO @metrics_hook: valid age_mae: 8.727 +-0.109 (17639) 2020-02-01 17:12:35.925: INFO @metrics_hook: valid gender_accuracy: 0.828 +-0.006 (17639) 2020-02-01 17:12:37.520: INFO @decay_lr : LR updated to `9.655207e-05` 2020-02-01 17:12:37.521: INFO @log_profile : T train: 122.195489 2020-02-01 17:12:37.521: INFO @log_profile : T valid: 5.501954 2020-02-01 17:12:37.521: INFO @log_profile : T read data: 2.853036 2020-02-01 17:12:37.521: INFO @log_profile : T hooks: 4.941726 2020-02-01 17:12:37.521: INFO @main_loop : Epoch 7 done 2020-02-01 17:12:37.521: INFO @main_loop : Training epoch 8 2020-02-01 17:14:48.614: INFO @log_variables: train loss nanmean: 1.343759 2020-02-01 17:14:48.614: INFO @log_variables: train age_loss mean: 9.639970 2020-02-01 17:14:48.614: INFO @log_variables: train gender_loss mean: 0.373591 2020-02-01 17:14:48.614: INFO @log_variables: train age_mae mean: 10.127967 2020-02-01 17:14:48.615: INFO @log_variables: train gender_accuracy mean: 0.822210 2020-02-01 17:14:48.615: INFO @log_variables: train gender_confidence/loss nanmean: 0.072081 2020-02-01 17:14:48.615: INFO @log_variables: train gender_confidence/accuracy mean: 0.730457 2020-02-01 17:14:48.615: INFO @log_variables: train age_confidence/loss mean: 0.055689 2020-02-01 17:14:48.615: INFO @log_variables: train age_confidence/accuracy mean: 0.656679 2020-02-01 17:14:48.615: INFO @log_variables: valid loss nanmean: 1.195528 2020-02-01 17:14:48.615: INFO @log_variables: valid age_loss mean: 8.059626 2020-02-01 17:14:48.615: INFO @log_variables: valid gender_loss mean: 0.367947 2020-02-01 17:14:48.615: INFO @log_variables: valid age_mae mean: 8.547208 2020-02-01 17:14:48.615: INFO @log_variables: valid gender_accuracy mean: 0.828845 2020-02-01 17:14:48.615: INFO @log_variables: valid gender_confidence/loss nanmean: 0.068414 2020-02-01 17:14:48.615: INFO @log_variables: valid gender_confidence/accuracy mean: 0.747435 2020-02-01 17:14:48.615: INFO @log_variables: valid age_confidence/loss mean: 0.059923 2020-02-01 17:14:48.615: INFO @log_variables: valid age_confidence/accuracy mean: 0.623731 2020-02-01 17:14:48.615: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:14:48.622: INFO @metrics_hook: train age_mae: 10.128 +-0.052 (110372) 2020-02-01 17:14:48.629: INFO @metrics_hook: train gender_accuracy: 0.822 +-0.002 (110372) 2020-02-01 17:14:51.396: INFO @metrics_hook: valid age_mae: 8.547 +-0.106 (17639) 2020-02-01 17:14:51.397: INFO @metrics_hook: valid gender_accuracy: 0.829 +-0.006 (17639) 2020-02-01 17:14:52.862: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:14:52.863: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.70 +- 0.22 2020-02-01 17:14:52.863: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.16 2020-02-01 17:14:52.863: INFO @evaluate_confidence: Average confidence of all samples 0.67 +- 0.23 2020-02-01 17:14:53.005: INFO @evaluate_confidence: Previous accuracy would be: 82.22 2020-02-01 17:14:53.005: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69] 2020-02-01 17:14:53.042: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [89.27, 89.67, 90.04, 90.41, 90.82, 91.21, 91.57, 91.9, 92.23, 92.53, 92.8, 93.1, 93.37, 93.63, 93.91, 94.19, 94.44, 94.7, 94.92, 95.15] 2020-02-01 17:14:53.042: INFO @evaluate_confidence: Dropped ratios are: [29.62, 31.18, 32.65, 34.01, 35.43, 36.81, 38.17, 39.53, 40.84, 42.09, 43.31, 44.55, 45.73, 46.93, 48.07, 49.22, 50.29, 51.41, 52.48, 53.57] 2020-02-01 17:14:53.090: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:14:53.090: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.43 +- 0.14 2020-02-01 17:14:53.090: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.11 2020-02-01 17:14:53.091: INFO @evaluate_confidence: Average confidence of all samples 0.40 +- 0.13 2020-02-01 17:14:53.230: INFO @evaluate_confidence: Previous accuracy would be: 34.11 2020-02-01 17:14:53.230: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43] 2020-02-01 17:14:53.240: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [41.9, 42.61, 43.37, 44.07, 44.83] 2020-02-01 17:14:53.240: INFO @evaluate_confidence: Dropped ratios are: [56.46, 59.89, 63.15, 66.02, 68.75] 2020-02-01 17:14:53.248: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:14:53.248: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.72 +- 0.22 2020-02-01 17:14:53.248: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.16 2020-02-01 17:14:53.248: INFO @evaluate_confidence: Average confidence of all samples 0.68 +- 0.23 2020-02-01 17:14:53.345: INFO @evaluate_confidence: Previous accuracy would be: 82.88 2020-02-01 17:14:53.345: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72] 2020-02-01 17:14:53.351: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [89.93, 90.15, 90.5, 90.74, 90.95, 91.19, 91.59, 91.92, 92.17, 92.4, 92.69, 93.06, 93.28, 93.56, 93.81, 94.04, 94.41, 94.61, 94.87, 95.08, 95.25, 95.51, 95.69] 2020-02-01 17:14:53.351: INFO @evaluate_confidence: Dropped ratios are: [27.85, 29.12, 30.44, 31.77, 33.09, 34.33, 35.63, 37.03, 38.19, 39.3, 40.41, 41.66, 42.77, 43.91, 44.86, 46.07, 47.07, 48.23, 49.35, 50.46, 51.52, 52.54, 53.6] 2020-02-01 17:14:53.359: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:14:53.359: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.39 +- 0.09 2020-02-01 17:14:53.359: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.08 2020-02-01 17:14:53.359: INFO @evaluate_confidence: Average confidence of all samples 0.37 +- 0.08 2020-02-01 17:14:53.484: INFO @evaluate_confidence: Previous accuracy would be: 37.97 2020-02-01 17:14:53.484: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38] 2020-02-01 17:14:53.485: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [45.15, 45.82, 46.34] 2020-02-01 17:14:53.485: INFO @evaluate_confidence: Dropped ratios are: [53.43, 58.34, 63.09] 2020-02-01 17:14:53.535: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:14:54.224: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:14:54.304: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:14:54.742: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:14:54.961: INFO @decay_lr : LR updated to `9.606931e-05` 2020-02-01 17:14:54.963: INFO @log_profile : T train: 122.076780 2020-02-01 17:14:54.963: INFO @log_profile : T valid: 5.487596 2020-02-01 17:14:54.963: INFO @log_profile : T read data: 2.836962 2020-02-01 17:14:54.963: INFO @log_profile : T hooks: 6.962062 2020-02-01 17:14:54.963: INFO @main_loop : Epoch 8 done 2020-02-01 17:14:54.963: INFO @main_loop : Training epoch 9 2020-02-01 17:17:05.127: INFO @log_variables: train loss nanmean: 1.311778 2020-02-01 17:17:05.127: INFO @log_variables: train age_loss mean: 9.421727 2020-02-01 17:17:05.128: INFO @log_variables: train gender_loss mean: 0.360271 2020-02-01 17:17:05.128: INFO @log_variables: train age_mae mean: 9.909142 2020-02-01 17:17:05.128: INFO @log_variables: train gender_accuracy mean: 0.833722 2020-02-01 17:17:05.128: INFO @log_variables: train gender_confidence/loss nanmean: 0.071853 2020-02-01 17:17:05.128: INFO @log_variables: train gender_confidence/accuracy mean: 0.735415 2020-02-01 17:17:05.128: INFO @log_variables: train age_confidence/loss mean: 0.055885 2020-02-01 17:17:05.128: INFO @log_variables: train age_confidence/accuracy mean: 0.655707 2020-02-01 17:17:05.128: INFO @log_variables: valid loss nanmean: 1.133864 2020-02-01 17:17:05.128: INFO @log_variables: valid age_loss mean: 7.855251 2020-02-01 17:17:05.128: INFO @log_variables: valid gender_loss mean: 0.321707 2020-02-01 17:17:05.128: INFO @log_variables: valid age_mae mean: 8.341433 2020-02-01 17:17:05.128: INFO @log_variables: valid gender_accuracy mean: 0.852032 2020-02-01 17:17:05.128: INFO @log_variables: valid gender_confidence/loss nanmean: 0.066721 2020-02-01 17:17:05.128: INFO @log_variables: valid gender_confidence/accuracy mean: 0.756165 2020-02-01 17:17:05.128: INFO @log_variables: valid age_confidence/loss mean: 0.060568 2020-02-01 17:17:05.128: INFO @log_variables: valid age_confidence/accuracy mean: 0.610806 2020-02-01 17:17:05.128: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:17:05.136: INFO @metrics_hook: train age_mae: 9.909 +-0.051 (110592) 2020-02-01 17:17:05.144: INFO @metrics_hook: train gender_accuracy: 0.834 +-0.002 (110592) 2020-02-01 17:17:07.923: INFO @metrics_hook: valid age_mae: 8.341 +-0.104 (17639) 2020-02-01 17:17:07.925: INFO @metrics_hook: valid gender_accuracy: 0.852 +-0.005 (17639) 2020-02-01 17:17:09.522: INFO @decay_lr : LR updated to `9.558896e-05` 2020-02-01 17:17:09.815: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:17:09.817: INFO @log_profile : T train: 122.029201 2020-02-01 17:17:09.817: INFO @log_profile : T valid: 5.522266 2020-02-01 17:17:09.817: INFO @log_profile : T read data: 1.911600 2020-02-01 17:17:09.817: INFO @log_profile : T hooks: 5.313589 2020-02-01 17:17:09.817: INFO @main_loop : Epoch 9 done 2020-02-01 17:17:09.817: INFO @main_loop : Training epoch 10 2020-02-01 17:19:28.259: INFO @log_variables: train loss nanmean: 1.276423 2020-02-01 17:19:28.260: INFO @log_variables: train age_loss mean: 9.143496 2020-02-01 17:19:28.260: INFO @log_variables: train gender_loss mean: 0.349278 2020-02-01 17:19:28.260: INFO @log_variables: train age_mae mean: 9.630597 2020-02-01 17:19:28.260: INFO @log_variables: train gender_accuracy mean: 0.838030 2020-02-01 17:19:28.260: INFO @log_variables: train gender_confidence/loss nanmean: 0.070937 2020-02-01 17:19:28.260: INFO @log_variables: train gender_confidence/accuracy mean: 0.741819 2020-02-01 17:19:28.260: INFO @log_variables: train age_confidence/loss mean: 0.056733 2020-02-01 17:19:28.260: INFO @log_variables: train age_confidence/accuracy mean: 0.649449 2020-02-01 17:19:28.260: INFO @log_variables: valid loss nanmean: 1.170828 2020-02-01 17:19:28.260: INFO @log_variables: valid age_loss mean: 7.979216 2020-02-01 17:19:28.260: INFO @log_variables: valid gender_loss mean: 0.351178 2020-02-01 17:19:28.260: INFO @log_variables: valid age_mae mean: 8.464720 2020-02-01 17:19:28.260: INFO @log_variables: valid gender_accuracy mean: 0.836045 2020-02-01 17:19:28.260: INFO @log_variables: valid gender_confidence/loss nanmean: 0.064840 2020-02-01 17:19:28.260: INFO @log_variables: valid gender_confidence/accuracy mean: 0.770565 2020-02-01 17:19:28.260: INFO @log_variables: valid age_confidence/loss mean: 0.061352 2020-02-01 17:19:28.260: INFO @log_variables: valid age_confidence/accuracy mean: 0.613243 2020-02-01 17:19:28.260: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:19:28.268: INFO @metrics_hook: train age_mae: 9.631 +-0.050 (110372) 2020-02-01 17:19:28.275: INFO @metrics_hook: train gender_accuracy: 0.838 +-0.002 (110372) 2020-02-01 17:19:31.560: INFO @metrics_hook: valid age_mae: 8.465 +-0.111 (17639) 2020-02-01 17:19:31.561: INFO @metrics_hook: valid gender_accuracy: 0.836 +-0.006 (17639) 2020-02-01 17:19:33.007: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:19:33.007: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.71 +- 0.22 2020-02-01 17:19:33.007: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.16 2020-02-01 17:19:33.008: INFO @evaluate_confidence: Average confidence of all samples 0.68 +- 0.23 2020-02-01 17:19:33.142: INFO @evaluate_confidence: Previous accuracy would be: 83.80 2020-02-01 17:19:33.143: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7] 2020-02-01 17:19:33.182: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [90.39, 90.77, 91.12, 91.44, 91.79, 92.15, 92.49, 92.81, 93.12, 93.4, 93.7, 94.02, 94.25, 94.47, 94.74, 94.96, 95.18, 95.39, 95.58, 95.81, 96.02, 96.2] 2020-02-01 17:19:33.182: INFO @evaluate_confidence: Dropped ratios are: [27.42, 28.88, 30.26, 31.64, 33.04, 34.4, 35.71, 36.99, 38.23, 39.51, 40.66, 41.79, 42.89, 44.01, 45.08, 46.14, 47.22, 48.21, 49.21, 50.27, 51.38, 52.38] 2020-02-01 17:19:33.230: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:19:33.231: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.45 +- 0.14 2020-02-01 17:19:33.231: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.11 2020-02-01 17:19:33.231: INFO @evaluate_confidence: Average confidence of all samples 0.42 +- 0.12 2020-02-01 17:19:33.366: INFO @evaluate_confidence: Previous accuracy would be: 36.29 2020-02-01 17:19:33.366: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43, 0.44] 2020-02-01 17:19:33.376: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [45.5, 46.23, 46.97, 47.69, 48.47] 2020-02-01 17:19:33.376: INFO @evaluate_confidence: Dropped ratios are: [54.16, 57.63, 60.91, 63.8, 66.51] 2020-02-01 17:19:33.384: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:19:33.384: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.74 +- 0.21 2020-02-01 17:19:33.384: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.15 2020-02-01 17:19:33.384: INFO @evaluate_confidence: Average confidence of all samples 0.70 +- 0.22 2020-02-01 17:19:33.483: INFO @evaluate_confidence: Previous accuracy would be: 83.60 2020-02-01 17:19:33.483: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:19:33.489: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [90.85, 91.18, 91.49, 91.87, 92.13, 92.37, 92.56, 92.91, 93.15, 93.36, 93.59, 93.89, 94.09, 94.36, 94.57, 94.76, 94.88, 95.04, 95.19, 95.38, 95.62, 95.84, 96.08, 96.37, 96.57] 2020-02-01 17:19:33.489: INFO @evaluate_confidence: Dropped ratios are: [25.74, 27.15, 28.57, 29.92, 31.31, 32.61, 33.97, 35.23, 36.37, 37.46, 38.55, 39.63, 40.69, 41.78, 42.73, 43.75, 44.8, 45.82, 46.94, 47.88, 48.85, 49.83, 50.86, 52.02, 53.06] 2020-02-01 17:19:33.497: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:19:33.497: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.43 +- 0.10 2020-02-01 17:19:33.497: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.08 2020-02-01 17:19:33.498: INFO @evaluate_confidence: Average confidence of all samples 0.41 +- 0.09 2020-02-01 17:19:33.624: INFO @evaluate_confidence: Previous accuracy would be: 40.30 2020-02-01 17:19:33.624: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43] 2020-02-01 17:19:33.626: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [47.38, 48.1, 48.69, 49.43, 50.41] 2020-02-01 17:19:33.626: INFO @evaluate_confidence: Dropped ratios are: [48.91, 53.63, 58.16, 62.41, 66.57] 2020-02-01 17:19:33.679: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:19:34.351: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:19:34.433: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:19:34.847: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:19:34.922: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:19:35.590: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:19:35.671: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 17:19:35.673: INFO @evaluate_gender-age_model: groups 0 9.839665 1 9.450242 2 8.643442 3 7.493108 4 8.281487 5 9.840569 6 11.693428 7 15.261400 Name: errors, dtype: float64 2020-02-01 17:19:35.674: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:19:36.088: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:19:36.148: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 17:19:36.149: INFO @evaluate_gender-age_model: groups 0 9.188713 1 7.829815 2 7.299475 3 6.150088 4 8.251206 5 8.269855 6 12.111769 7 20.297691 Name: errors, dtype: float64 2020-02-01 17:19:36.295: INFO @decay_lr : LR updated to `9.511102e-05` 2020-02-01 17:19:36.296: INFO @log_profile : T train: 127.668703 2020-02-01 17:19:36.296: INFO @log_profile : T valid: 5.579508 2020-02-01 17:19:36.296: INFO @log_profile : T read data: 2.844587 2020-02-01 17:19:36.296: INFO @log_profile : T hooks: 10.310410 2020-02-01 17:19:36.296: INFO @main_loop : Epoch 10 done 2020-02-01 17:19:36.296: INFO @main_loop : Training epoch 11 2020-02-01 17:21:46.919: INFO @log_variables: train loss nanmean: 1.243951 2020-02-01 17:21:46.919: INFO @log_variables: train age_loss mean: 8.959956 2020-02-01 17:21:46.920: INFO @log_variables: train gender_loss mean: 0.332468 2020-02-01 17:21:46.920: INFO @log_variables: train age_mae mean: 9.446294 2020-02-01 17:21:46.920: INFO @log_variables: train gender_accuracy mean: 0.847851 2020-02-01 17:21:46.920: INFO @log_variables: train gender_confidence/loss nanmean: 0.070165 2020-02-01 17:21:46.920: INFO @log_variables: train gender_confidence/accuracy mean: 0.747626 2020-02-01 17:21:46.920: INFO @log_variables: train age_confidence/loss mean: 0.057000 2020-02-01 17:21:46.920: INFO @log_variables: train age_confidence/accuracy mean: 0.646595 2020-02-01 17:21:46.920: INFO @log_variables: valid loss nanmean: 1.117641 2020-02-01 17:21:46.920: INFO @log_variables: valid age_loss mean: 7.777112 2020-02-01 17:21:46.920: INFO @log_variables: valid gender_loss mean: 0.314671 2020-02-01 17:21:46.920: INFO @log_variables: valid age_mae mean: 8.263027 2020-02-01 17:21:46.920: INFO @log_variables: valid gender_accuracy mean: 0.855774 2020-02-01 17:21:46.920: INFO @log_variables: valid gender_confidence/loss nanmean: 0.063119 2020-02-01 17:21:46.920: INFO @log_variables: valid gender_confidence/accuracy mean: 0.769205 2020-02-01 17:21:46.920: INFO @log_variables: valid age_confidence/loss mean: 0.061447 2020-02-01 17:21:46.920: INFO @log_variables: valid age_confidence/accuracy mean: 0.608028 2020-02-01 17:21:46.920: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:21:46.928: INFO @metrics_hook: train age_mae: 9.446 +-0.050 (110372) 2020-02-01 17:21:46.935: INFO @metrics_hook: train gender_accuracy: 0.848 +-0.002 (110372) 2020-02-01 17:21:49.670: INFO @metrics_hook: valid age_mae: 8.263 +-0.107 (17639) 2020-02-01 17:21:49.671: INFO @metrics_hook: valid gender_accuracy: 0.856 +-0.005 (17639) 2020-02-01 17:21:51.262: INFO @decay_lr : LR updated to `9.463546e-05` 2020-02-01 17:21:51.547: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:21:51.549: INFO @log_profile : T train: 121.621379 2020-02-01 17:21:51.549: INFO @log_profile : T valid: 5.491041 2020-02-01 17:21:51.549: INFO @log_profile : T read data: 2.813658 2020-02-01 17:21:51.549: INFO @log_profile : T hooks: 5.251782 2020-02-01 17:21:51.549: INFO @main_loop : Epoch 11 done 2020-02-01 17:21:51.549: INFO @main_loop : Training epoch 12 2020-02-01 17:24:01.120: INFO @log_variables: train loss nanmean: 1.220773 2020-02-01 17:24:01.120: INFO @log_variables: train age_loss mean: 8.776491 2020-02-01 17:24:01.120: INFO @log_variables: train gender_loss mean: 0.325525 2020-02-01 17:24:01.121: INFO @log_variables: train age_mae mean: 9.262636 2020-02-01 17:24:01.121: INFO @log_variables: train gender_accuracy mean: 0.851373 2020-02-01 17:24:01.121: INFO @log_variables: train gender_confidence/loss nanmean: 0.069431 2020-02-01 17:24:01.121: INFO @log_variables: train gender_confidence/accuracy mean: 0.751591 2020-02-01 17:24:01.121: INFO @log_variables: train age_confidence/loss mean: 0.057548 2020-02-01 17:24:01.121: INFO @log_variables: train age_confidence/accuracy mean: 0.642361 2020-02-01 17:24:01.121: INFO @log_variables: valid loss nanmean: 1.063935 2020-02-01 17:24:01.121: INFO @log_variables: valid age_loss mean: 7.283786 2020-02-01 17:24:01.121: INFO @log_variables: valid gender_loss mean: 0.300368 2020-02-01 17:24:01.121: INFO @log_variables: valid age_mae mean: 7.767942 2020-02-01 17:24:01.121: INFO @log_variables: valid gender_accuracy mean: 0.861613 2020-02-01 17:24:01.121: INFO @log_variables: valid gender_confidence/loss nanmean: 0.063722 2020-02-01 17:24:01.121: INFO @log_variables: valid gender_confidence/accuracy mean: 0.780883 2020-02-01 17:24:01.121: INFO @log_variables: valid age_confidence/loss mean: 0.064988 2020-02-01 17:24:01.121: INFO @log_variables: valid age_confidence/accuracy mean: 0.575373 2020-02-01 17:24:01.121: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:24:01.128: INFO @metrics_hook: train age_mae: 9.263 +-0.049 (110592) 2020-02-01 17:24:01.135: INFO @metrics_hook: train gender_accuracy: 0.851 +-0.002 (110592) 2020-02-01 17:24:04.564: INFO @metrics_hook: valid age_mae: 7.768 +-0.104 (17639) 2020-02-01 17:24:04.565: INFO @metrics_hook: valid gender_accuracy: 0.862 +-0.005 (17639) 2020-02-01 17:24:06.011: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:24:06.012: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.73 +- 0.23 2020-02-01 17:24:06.012: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.16 2020-02-01 17:24:06.012: INFO @evaluate_confidence: Average confidence of all samples 0.69 +- 0.24 2020-02-01 17:24:06.156: INFO @evaluate_confidence: Previous accuracy would be: 85.14 2020-02-01 17:24:06.156: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73] 2020-02-01 17:24:06.203: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [91.54, 91.84, 92.14, 92.42, 92.73, 93.03, 93.28, 93.54, 93.82, 94.06, 94.32, 94.56, 94.78, 94.99, 95.21, 95.4, 95.61, 95.79, 95.96, 96.14, 96.29, 96.45, 96.63, 96.79, 96.95, 97.11] 2020-02-01 17:24:06.203: INFO @evaluate_confidence: Dropped ratios are: [26.03, 27.24, 28.46, 29.7, 30.95, 32.14, 33.26, 34.36, 35.49, 36.57, 37.67, 38.73, 39.8, 40.84, 41.84, 42.84, 43.83, 44.81, 45.76, 46.77, 47.71, 48.58, 49.51, 50.45, 51.39, 52.31] 2020-02-01 17:24:06.253: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:24:06.253: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.47 +- 0.14 2020-02-01 17:24:06.253: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.11 2020-02-01 17:24:06.253: INFO @evaluate_confidence: Average confidence of all samples 0.43 +- 0.13 2020-02-01 17:24:06.389: INFO @evaluate_confidence: Previous accuracy would be: 38.04 2020-02-01 17:24:06.389: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47] 2020-02-01 17:24:06.404: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [48.15, 48.9, 49.6, 50.32, 51.02, 51.77, 52.47] 2020-02-01 17:24:06.404: INFO @evaluate_confidence: Dropped ratios are: [53.16, 56.25, 59.31, 62.19, 64.87, 67.37, 69.76] 2020-02-01 17:24:06.412: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:24:06.412: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.74 +- 0.22 2020-02-01 17:24:06.412: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.15 2020-02-01 17:24:06.413: INFO @evaluate_confidence: Average confidence of all samples 0.71 +- 0.23 2020-02-01 17:24:06.511: INFO @evaluate_confidence: Previous accuracy would be: 86.16 2020-02-01 17:24:06.512: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:24:06.518: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [92.51, 92.81, 93.14, 93.46, 93.7, 93.98, 94.2, 94.45, 94.72, 94.88, 95.09, 95.26, 95.51, 95.68, 95.8, 95.93, 96.14, 96.29, 96.4, 96.47, 96.58, 96.72, 96.86, 97.07, 97.25, 97.47, 97.61, 97.8] 2020-02-01 17:24:06.518: INFO @evaluate_confidence: Dropped ratios are: [22.61, 23.81, 25.01, 26.09, 27.07, 28.15, 29.28, 30.38, 31.45, 32.5, 33.6, 34.67, 35.77, 36.8, 37.72, 38.66, 39.71, 40.63, 41.58, 42.53, 43.56, 44.59, 45.67, 46.61, 47.52, 48.55, 49.53, 50.43] 2020-02-01 17:24:06.526: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:24:06.526: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.42 +- 0.09 2020-02-01 17:24:06.526: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.08 2020-02-01 17:24:06.526: INFO @evaluate_confidence: Average confidence of all samples 0.40 +- 0.09 2020-02-01 17:24:06.651: INFO @evaluate_confidence: Previous accuracy would be: 44.24 2020-02-01 17:24:06.651: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41] 2020-02-01 17:24:06.653: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [52.0, 52.15, 52.49, 52.65] 2020-02-01 17:24:06.653: INFO @evaluate_confidence: Dropped ratios are: [48.76, 53.15, 57.08, 61.19] 2020-02-01 17:24:06.706: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:24:07.390: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:24:07.471: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:24:07.920: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:24:08.137: INFO @decay_lr : LR updated to `9.4162286e-05` 2020-02-01 17:24:08.440: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:24:08.443: INFO @log_profile : T train: 121.517843 2020-02-01 17:24:08.443: INFO @log_profile : T valid: 5.489935 2020-02-01 17:24:08.443: INFO @log_profile : T read data: 1.866988 2020-02-01 17:24:08.443: INFO @log_profile : T hooks: 7.940831 2020-02-01 17:24:08.443: INFO @main_loop : Epoch 12 done 2020-02-01 17:24:08.443: INFO @main_loop : Training epoch 13 2020-02-01 17:26:19.113: INFO @log_variables: train loss nanmean: 1.189568 2020-02-01 17:26:19.113: INFO @log_variables: train age_loss mean: 8.557080 2020-02-01 17:26:19.113: INFO @log_variables: train gender_loss mean: 0.312543 2020-02-01 17:26:19.113: INFO @log_variables: train age_mae mean: 9.043044 2020-02-01 17:26:19.113: INFO @log_variables: train gender_accuracy mean: 0.858542 2020-02-01 17:26:19.113: INFO @log_variables: train gender_confidence/loss nanmean: 0.069339 2020-02-01 17:26:19.113: INFO @log_variables: train gender_confidence/accuracy mean: 0.756904 2020-02-01 17:26:19.113: INFO @log_variables: train age_confidence/loss mean: 0.058183 2020-02-01 17:26:19.113: INFO @log_variables: train age_confidence/accuracy mean: 0.637172 2020-02-01 17:26:19.113: INFO @log_variables: valid loss nanmean: 1.089216 2020-02-01 17:26:19.113: INFO @log_variables: valid age_loss mean: 7.675829 2020-02-01 17:26:19.113: INFO @log_variables: valid gender_loss mean: 0.292876 2020-02-01 17:26:19.113: INFO @log_variables: valid age_mae mean: 8.160729 2020-02-01 17:26:19.114: INFO @log_variables: valid gender_accuracy mean: 0.864108 2020-02-01 17:26:19.114: INFO @log_variables: valid gender_confidence/loss nanmean: 0.061900 2020-02-01 17:26:19.114: INFO @log_variables: valid gender_confidence/accuracy mean: 0.798288 2020-02-01 17:26:19.114: INFO @log_variables: valid age_confidence/loss mean: 0.063262 2020-02-01 17:26:19.114: INFO @log_variables: valid age_confidence/accuracy mean: 0.583707 2020-02-01 17:26:19.114: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:26:19.121: INFO @metrics_hook: train age_mae: 9.043 +-0.048 (110372) 2020-02-01 17:26:19.128: INFO @metrics_hook: train gender_accuracy: 0.859 +-0.002 (110372) 2020-02-01 17:26:21.960: INFO @metrics_hook: valid age_mae: 8.161 +-0.109 (17639) 2020-02-01 17:26:21.962: INFO @metrics_hook: valid gender_accuracy: 0.864 +-0.005 (17639) 2020-02-01 17:26:23.602: INFO @decay_lr : LR updated to `9.369147e-05` 2020-02-01 17:26:23.905: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:26:23.907: INFO @log_profile : T train: 121.643590 2020-02-01 17:26:23.907: INFO @log_profile : T valid: 5.488588 2020-02-01 17:26:23.908: INFO @log_profile : T read data: 2.838621 2020-02-01 17:26:23.908: INFO @log_profile : T hooks: 5.417351 2020-02-01 17:26:23.908: INFO @main_loop : Epoch 13 done 2020-02-01 17:26:23.908: INFO @main_loop : Training epoch 14 2020-02-01 17:28:34.297: INFO @log_variables: train loss nanmean: 1.169936 2020-02-01 17:28:34.297: INFO @log_variables: train age_loss mean: 8.400031 2020-02-01 17:28:34.297: INFO @log_variables: train gender_loss mean: 0.307176 2020-02-01 17:28:34.297: INFO @log_variables: train age_mae mean: 8.885638 2020-02-01 17:28:34.297: INFO @log_variables: train gender_accuracy mean: 0.861287 2020-02-01 17:28:34.297: INFO @log_variables: train gender_confidence/loss nanmean: 0.068482 2020-02-01 17:28:34.297: INFO @log_variables: train gender_confidence/accuracy mean: 0.759704 2020-02-01 17:28:34.298: INFO @log_variables: train age_confidence/loss mean: 0.058564 2020-02-01 17:28:34.298: INFO @log_variables: train age_confidence/accuracy mean: 0.634409 2020-02-01 17:28:34.298: INFO @log_variables: valid loss nanmean: 1.033912 2020-02-01 17:28:34.298: INFO @log_variables: valid age_loss mean: 7.145343 2020-02-01 17:28:34.298: INFO @log_variables: valid gender_loss mean: 0.283424 2020-02-01 17:28:34.298: INFO @log_variables: valid age_mae mean: 7.628684 2020-02-01 17:28:34.298: INFO @log_variables: valid gender_accuracy mean: 0.872215 2020-02-01 17:28:34.298: INFO @log_variables: valid gender_confidence/loss nanmean: 0.061371 2020-02-01 17:28:34.298: INFO @log_variables: valid gender_confidence/accuracy mean: 0.784171 2020-02-01 17:28:34.298: INFO @log_variables: valid age_confidence/loss mean: 0.065305 2020-02-01 17:28:34.298: INFO @log_variables: valid age_confidence/accuracy mean: 0.578774 2020-02-01 17:28:34.298: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:28:34.305: INFO @metrics_hook: train age_mae: 8.886 +-0.048 (110372) 2020-02-01 17:28:34.312: INFO @metrics_hook: train gender_accuracy: 0.861 +-0.002 (110372) 2020-02-01 17:28:37.040: INFO @metrics_hook: valid age_mae: 7.629 +-0.104 (17639) 2020-02-01 17:28:37.041: INFO @metrics_hook: valid gender_accuracy: 0.872 +-0.005 (17639) 2020-02-01 17:28:38.493: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:28:38.494: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.73 +- 0.23 2020-02-01 17:28:38.494: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.16 2020-02-01 17:28:38.494: INFO @evaluate_confidence: Average confidence of all samples 0.70 +- 0.24 2020-02-01 17:28:38.636: INFO @evaluate_confidence: Previous accuracy would be: 86.13 2020-02-01 17:28:38.636: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73] 2020-02-01 17:28:38.683: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [92.43, 92.75, 93.03, 93.31, 93.57, 93.84, 94.08, 94.34, 94.55, 94.77, 94.96, 95.16, 95.35, 95.57, 95.79, 95.98, 96.17, 96.35, 96.5, 96.63, 96.78, 96.95, 97.11, 97.24, 97.32, 97.47] 2020-02-01 17:28:38.683: INFO @evaluate_confidence: Dropped ratios are: [25.5, 26.68, 27.85, 29.04, 30.16, 31.26, 32.3, 33.36, 34.43, 35.48, 36.46, 37.45, 38.4, 39.4, 40.38, 41.33, 42.24, 43.17, 44.08, 44.99, 45.9, 46.83, 47.8, 48.71, 49.66, 50.58] 2020-02-01 17:28:38.730: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:28:38.731: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.48 +- 0.14 2020-02-01 17:28:38.731: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.11 2020-02-01 17:28:38.731: INFO @evaluate_confidence: Average confidence of all samples 0.44 +- 0.13 2020-02-01 17:28:38.867: INFO @evaluate_confidence: Previous accuracy would be: 39.49 2020-02-01 17:28:38.867: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47] 2020-02-01 17:28:38.882: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [48.94, 49.7, 50.39, 51.22, 51.94, 52.58, 53.36] 2020-02-01 17:28:38.882: INFO @evaluate_confidence: Dropped ratios are: [49.49, 52.67, 55.82, 58.75, 61.58, 64.24, 66.75] 2020-02-01 17:28:38.889: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:28:38.890: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.75 +- 0.23 2020-02-01 17:28:38.890: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.15 2020-02-01 17:28:38.890: INFO @evaluate_confidence: Average confidence of all samples 0.71 +- 0.24 2020-02-01 17:28:38.991: INFO @evaluate_confidence: Previous accuracy would be: 87.22 2020-02-01 17:28:38.991: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:28:38.998: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.42, 93.71, 93.9, 94.09, 94.29, 94.51, 94.71, 94.92, 95.13, 95.26, 95.55, 95.73, 95.92, 96.08, 96.3, 96.46, 96.7, 96.92, 97.04, 97.14, 97.21, 97.35, 97.47, 97.62, 97.74, 97.91, 98.08, 98.2, 98.3] 2020-02-01 17:28:38.998: INFO @evaluate_confidence: Dropped ratios are: [21.54, 22.82, 23.95, 24.93, 25.9, 26.84, 27.95, 29.09, 30.1, 31.02, 32.04, 32.96, 33.92, 34.88, 35.84, 36.76, 37.69, 38.6, 39.62, 40.47, 41.24, 42.09, 42.89, 43.78, 44.75, 45.81, 46.82, 47.71, 48.52] 2020-02-01 17:28:39.006: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:28:39.006: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.43 +- 0.11 2020-02-01 17:28:39.006: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.09 2020-02-01 17:28:39.006: INFO @evaluate_confidence: Average confidence of all samples 0.41 +- 0.10 2020-02-01 17:28:39.130: INFO @evaluate_confidence: Previous accuracy would be: 45.13 2020-02-01 17:28:39.130: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43] 2020-02-01 17:28:39.131: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [54.12, 54.49, 55.2, 55.52, 55.64] 2020-02-01 17:28:39.131: INFO @evaluate_confidence: Dropped ratios are: [49.72, 53.31, 57.18, 60.8, 64.24] 2020-02-01 17:28:39.183: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:28:39.874: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:28:39.955: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:28:40.399: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:28:40.617: INFO @decay_lr : LR updated to `9.322302e-05` 2020-02-01 17:28:40.909: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:28:40.912: INFO @log_profile : T train: 121.423982 2020-02-01 17:28:40.912: INFO @log_profile : T valid: 5.490980 2020-02-01 17:28:40.912: INFO @log_profile : T read data: 2.810649 2020-02-01 17:28:40.912: INFO @log_profile : T hooks: 7.200560 2020-02-01 17:28:40.912: INFO @main_loop : Epoch 14 done 2020-02-01 17:28:40.912: INFO @main_loop : Training epoch 15 2020-02-01 17:30:50.888: INFO @log_variables: train loss nanmean: 1.141867 2020-02-01 17:30:50.888: INFO @log_variables: train age_loss mean: 8.197689 2020-02-01 17:30:50.888: INFO @log_variables: train gender_loss mean: 0.296986 2020-02-01 17:30:50.888: INFO @log_variables: train age_mae mean: 8.682297 2020-02-01 17:30:50.888: INFO @log_variables: train gender_accuracy mean: 0.867314 2020-02-01 17:30:50.889: INFO @log_variables: train gender_confidence/loss nanmean: 0.067869 2020-02-01 17:30:50.889: INFO @log_variables: train gender_confidence/accuracy mean: 0.767262 2020-02-01 17:30:50.889: INFO @log_variables: train age_confidence/loss mean: 0.058766 2020-02-01 17:30:50.889: INFO @log_variables: train age_confidence/accuracy mean: 0.634069 2020-02-01 17:30:50.889: INFO @log_variables: valid loss nanmean: 1.034667 2020-02-01 17:30:50.889: INFO @log_variables: valid age_loss mean: 7.148405 2020-02-01 17:30:50.889: INFO @log_variables: valid gender_loss mean: 0.283862 2020-02-01 17:30:50.889: INFO @log_variables: valid age_mae mean: 7.632762 2020-02-01 17:30:50.889: INFO @log_variables: valid gender_accuracy mean: 0.874936 2020-02-01 17:30:50.889: INFO @log_variables: valid gender_confidence/loss nanmean: 0.062601 2020-02-01 17:30:50.889: INFO @log_variables: valid gender_confidence/accuracy mean: 0.798741 2020-02-01 17:30:50.889: INFO @log_variables: valid age_confidence/loss mean: 0.064155 2020-02-01 17:30:50.889: INFO @log_variables: valid age_confidence/accuracy mean: 0.587108 2020-02-01 17:30:50.890: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:30:50.897: INFO @metrics_hook: train age_mae: 8.682 +-0.047 (110592) 2020-02-01 17:30:50.904: INFO @metrics_hook: train gender_accuracy: 0.867 +-0.002 (110592) 2020-02-01 17:30:54.622: INFO @metrics_hook: valid age_mae: 7.633 +-0.103 (17639) 2020-02-01 17:30:54.623: INFO @metrics_hook: valid gender_accuracy: 0.875 +-0.005 (17639) 2020-02-01 17:30:56.298: INFO @decay_lr : LR updated to `9.2756905e-05` 2020-02-01 17:30:56.602: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:30:56.605: INFO @log_profile : T train: 121.880003 2020-02-01 17:30:56.605: INFO @log_profile : T valid: 5.495562 2020-02-01 17:30:56.605: INFO @log_profile : T read data: 1.920870 2020-02-01 17:30:56.605: INFO @log_profile : T hooks: 6.317750 2020-02-01 17:30:56.605: INFO @main_loop : Epoch 15 done 2020-02-01 17:30:56.605: INFO @main_loop : Training epoch 16 2020-02-01 17:33:16.058: INFO @log_variables: train loss nanmean: 1.128174 2020-02-01 17:33:16.058: INFO @log_variables: train age_loss mean: 8.103747 2020-02-01 17:33:16.058: INFO @log_variables: train gender_loss mean: 0.291337 2020-02-01 17:33:16.058: INFO @log_variables: train age_mae mean: 8.588318 2020-02-01 17:33:16.058: INFO @log_variables: train gender_accuracy mean: 0.869931 2020-02-01 17:33:16.058: INFO @log_variables: train gender_confidence/loss nanmean: 0.067447 2020-02-01 17:33:16.058: INFO @log_variables: train gender_confidence/accuracy mean: 0.771038 2020-02-01 17:33:16.058: INFO @log_variables: train age_confidence/loss mean: 0.059172 2020-02-01 17:33:16.058: INFO @log_variables: train age_confidence/accuracy mean: 0.632307 2020-02-01 17:33:16.058: INFO @log_variables: valid loss nanmean: 1.025516 2020-02-01 17:33:16.058: INFO @log_variables: valid age_loss mean: 7.136369 2020-02-01 17:33:16.058: INFO @log_variables: valid gender_loss mean: 0.274226 2020-02-01 17:33:16.058: INFO @log_variables: valid age_mae mean: 7.620796 2020-02-01 17:33:16.058: INFO @log_variables: valid gender_accuracy mean: 0.877771 2020-02-01 17:33:16.058: INFO @log_variables: valid gender_confidence/loss nanmean: 0.063196 2020-02-01 17:33:16.059: INFO @log_variables: valid gender_confidence/accuracy mean: 0.781904 2020-02-01 17:33:16.059: INFO @log_variables: valid age_confidence/loss mean: 0.064263 2020-02-01 17:33:16.059: INFO @log_variables: valid age_confidence/accuracy mean: 0.576847 2020-02-01 17:33:16.059: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:33:16.066: INFO @metrics_hook: train age_mae: 8.588 +-0.046 (110372) 2020-02-01 17:33:16.073: INFO @metrics_hook: train gender_accuracy: 0.870 +-0.002 (110372) 2020-02-01 17:33:18.808: INFO @metrics_hook: valid age_mae: 7.621 +-0.102 (17639) 2020-02-01 17:33:18.810: INFO @metrics_hook: valid gender_accuracy: 0.878 +-0.005 (17639) 2020-02-01 17:33:20.262: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:33:20.262: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.74 +- 0.23 2020-02-01 17:33:20.262: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.17 2020-02-01 17:33:20.262: INFO @evaluate_confidence: Average confidence of all samples 0.71 +- 0.24 2020-02-01 17:33:20.405: INFO @evaluate_confidence: Previous accuracy would be: 86.99 2020-02-01 17:33:20.405: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:33:20.453: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.1, 93.4, 93.64, 93.89, 94.13, 94.36, 94.57, 94.8, 94.99, 95.19, 95.41, 95.59, 95.79, 95.96, 96.13, 96.28, 96.45, 96.6, 96.73, 96.85, 96.96, 97.07, 97.22, 97.35, 97.47, 97.59, 97.68, 97.79] 2020-02-01 17:33:20.453: INFO @evaluate_confidence: Dropped ratios are: [23.71, 24.83, 25.91, 26.98, 28.03, 29.09, 30.1, 31.1, 32.08, 33.06, 34.07, 35.01, 35.96, 36.89, 37.8, 38.68, 39.6, 40.5, 41.37, 42.23, 43.11, 43.98, 44.88, 45.8, 46.67, 47.55, 48.41, 49.26] 2020-02-01 17:33:20.501: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:33:20.501: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.15 2020-02-01 17:33:20.501: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.11 2020-02-01 17:33:20.501: INFO @evaluate_confidence: Average confidence of all samples 0.45 +- 0.13 2020-02-01 17:33:20.635: INFO @evaluate_confidence: Previous accuracy would be: 40.87 2020-02-01 17:33:20.636: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 17:33:20.652: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [50.98, 51.68, 52.38, 53.15, 53.93, 54.68, 55.42, 56.24] 2020-02-01 17:33:20.652: INFO @evaluate_confidence: Dropped ratios are: [50.12, 53.14, 56.12, 59.0, 61.74, 64.36, 66.83, 69.22] 2020-02-01 17:33:20.660: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:33:20.660: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.74 +- 0.22 2020-02-01 17:33:20.660: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.15 2020-02-01 17:33:20.660: INFO @evaluate_confidence: Average confidence of all samples 0.70 +- 0.23 2020-02-01 17:33:20.758: INFO @evaluate_confidence: Previous accuracy would be: 87.78 2020-02-01 17:33:20.758: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:33:20.765: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.61, 93.74, 94.04, 94.31, 94.59, 94.82, 95.05, 95.28, 95.42, 95.65, 95.86, 96.08, 96.27, 96.44, 96.63, 96.79, 96.96, 97.12, 97.25, 97.4, 97.57, 97.69, 97.79, 97.88, 98.05, 98.19, 98.32, 98.47, 98.53] 2020-02-01 17:33:20.765: INFO @evaluate_confidence: Dropped ratios are: [21.68, 22.62, 23.76, 24.85, 26.03, 27.17, 28.15, 29.17, 30.17, 31.11, 32.16, 33.26, 34.22, 35.2, 36.11, 37.11, 38.06, 39.04, 40.09, 41.07, 42.08, 43.23, 44.03, 45.0, 46.01, 47.0, 48.0, 49.03, 50.01] 2020-02-01 17:33:20.773: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:33:20.773: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.43 +- 0.11 2020-02-01 17:33:20.773: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.09 2020-02-01 17:33:20.773: INFO @evaluate_confidence: Average confidence of all samples 0.41 +- 0.10 2020-02-01 17:33:20.898: INFO @evaluate_confidence: Previous accuracy would be: 43.91 2020-02-01 17:33:20.899: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43] 2020-02-01 17:33:20.900: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [52.19, 52.5, 53.06, 53.66, 53.81] 2020-02-01 17:33:20.900: INFO @evaluate_confidence: Dropped ratios are: [52.11, 56.08, 59.92, 63.58, 66.87] 2020-02-01 17:33:20.950: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:33:21.631: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:33:21.712: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:33:22.159: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:33:22.383: INFO @decay_lr : LR updated to `9.229312e-05` 2020-02-01 17:33:22.687: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:33:22.689: INFO @log_profile : T train: 129.647860 2020-02-01 17:33:22.689: INFO @log_profile : T valid: 6.295101 2020-02-01 17:33:22.689: INFO @log_profile : T read data: 2.822138 2020-02-01 17:33:22.689: INFO @log_profile : T hooks: 7.244234 2020-02-01 17:33:22.689: INFO @main_loop : Epoch 16 done 2020-02-01 17:33:22.690: INFO @main_loop : Training epoch 17 2020-02-01 17:35:39.966: INFO @log_variables: train loss nanmean: 1.105557 2020-02-01 17:35:39.966: INFO @log_variables: train age_loss mean: 7.916615 2020-02-01 17:35:39.966: INFO @log_variables: train gender_loss mean: 0.284907 2020-02-01 17:35:39.966: INFO @log_variables: train age_mae mean: 8.400811 2020-02-01 17:35:39.966: INFO @log_variables: train gender_accuracy mean: 0.873337 2020-02-01 17:35:39.967: INFO @log_variables: train gender_confidence/loss nanmean: 0.067125 2020-02-01 17:35:39.967: INFO @log_variables: train gender_confidence/accuracy mean: 0.771129 2020-02-01 17:35:39.967: INFO @log_variables: train age_confidence/loss mean: 0.059733 2020-02-01 17:35:39.967: INFO @log_variables: train age_confidence/accuracy mean: 0.629154 2020-02-01 17:35:39.967: INFO @log_variables: valid loss nanmean: 1.036123 2020-02-01 17:35:39.967: INFO @log_variables: valid age_loss mean: 6.974990 2020-02-01 17:35:39.967: INFO @log_variables: valid gender_loss mean: 0.301501 2020-02-01 17:35:39.967: INFO @log_variables: valid age_mae mean: 7.457836 2020-02-01 17:35:39.967: INFO @log_variables: valid gender_accuracy mean: 0.866432 2020-02-01 17:35:39.967: INFO @log_variables: valid gender_confidence/loss nanmean: 0.062810 2020-02-01 17:35:39.967: INFO @log_variables: valid gender_confidence/accuracy mean: 0.795510 2020-02-01 17:35:39.967: INFO @log_variables: valid age_confidence/loss mean: 0.065132 2020-02-01 17:35:39.967: INFO @log_variables: valid age_confidence/accuracy mean: 0.579795 2020-02-01 17:35:39.967: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:35:39.975: INFO @metrics_hook: train age_mae: 8.401 +-0.046 (110372) 2020-02-01 17:35:39.982: INFO @metrics_hook: train gender_accuracy: 0.873 +-0.002 (110372) 2020-02-01 17:35:42.748: INFO @metrics_hook: valid age_mae: 7.458 +-0.101 (17639) 2020-02-01 17:35:42.750: INFO @metrics_hook: valid gender_accuracy: 0.866 +-0.005 (17639) 2020-02-01 17:35:44.377: INFO @decay_lr : LR updated to `9.183166e-05` 2020-02-01 17:35:44.379: INFO @log_profile : T train: 128.170875 2020-02-01 17:35:44.379: INFO @log_profile : T valid: 5.549081 2020-02-01 17:35:44.379: INFO @log_profile : T read data: 2.865444 2020-02-01 17:35:44.379: INFO @log_profile : T hooks: 5.028686 2020-02-01 17:35:44.379: INFO @main_loop : Epoch 17 done 2020-02-01 17:35:44.379: INFO @main_loop : Training epoch 18 2020-02-01 17:37:54.383: INFO @log_variables: train loss nanmean: 1.090951 2020-02-01 17:37:54.383: INFO @log_variables: train age_loss mean: 7.852879 2020-02-01 17:37:54.383: INFO @log_variables: train gender_loss mean: 0.274993 2020-02-01 17:37:54.383: INFO @log_variables: train age_mae mean: 8.337028 2020-02-01 17:37:54.383: INFO @log_variables: train gender_accuracy mean: 0.879024 2020-02-01 17:37:54.383: INFO @log_variables: train gender_confidence/loss nanmean: 0.067014 2020-02-01 17:37:54.383: INFO @log_variables: train gender_confidence/accuracy mean: 0.774197 2020-02-01 17:37:54.383: INFO @log_variables: train age_confidence/loss mean: 0.060045 2020-02-01 17:37:54.383: INFO @log_variables: train age_confidence/accuracy mean: 0.627423 2020-02-01 17:37:54.383: INFO @log_variables: valid loss nanmean: 1.013097 2020-02-01 17:37:54.383: INFO @log_variables: valid age_loss mean: 7.033069 2020-02-01 17:37:54.383: INFO @log_variables: valid gender_loss mean: 0.272840 2020-02-01 17:37:54.383: INFO @log_variables: valid age_mae mean: 7.516203 2020-02-01 17:37:54.383: INFO @log_variables: valid gender_accuracy mean: 0.876467 2020-02-01 17:37:54.383: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059294 2020-02-01 17:37:54.384: INFO @log_variables: valid gender_confidence/accuracy mean: 0.795963 2020-02-01 17:37:54.384: INFO @log_variables: valid age_confidence/loss mean: 0.066397 2020-02-01 17:37:54.384: INFO @log_variables: valid age_confidence/accuracy mean: 0.556154 2020-02-01 17:37:54.384: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:37:54.391: INFO @metrics_hook: train age_mae: 8.337 +-0.045 (110592) 2020-02-01 17:37:54.398: INFO @metrics_hook: train gender_accuracy: 0.879 +-0.002 (110592) 2020-02-01 17:37:57.186: INFO @metrics_hook: valid age_mae: 7.516 +-0.104 (17639) 2020-02-01 17:37:57.187: INFO @metrics_hook: valid gender_accuracy: 0.876 +-0.005 (17639) 2020-02-01 17:37:58.639: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:37:58.639: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.75 +- 0.23 2020-02-01 17:37:58.639: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.17 2020-02-01 17:37:58.640: INFO @evaluate_confidence: Average confidence of all samples 0.71 +- 0.25 2020-02-01 17:37:58.779: INFO @evaluate_confidence: Previous accuracy would be: 87.90 2020-02-01 17:37:58.779: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:37:58.828: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.4, 93.64, 93.87, 94.09, 94.3, 94.53, 94.75, 94.96, 95.18, 95.36, 95.53, 95.69, 95.85, 96.01, 96.17, 96.36, 96.5, 96.66, 96.82, 96.95, 97.07, 97.21, 97.33, 97.47, 97.58, 97.67, 97.78, 97.88, 97.99] 2020-02-01 17:37:58.828: INFO @evaluate_confidence: Dropped ratios are: [22.19, 23.25, 24.27, 25.3, 26.27, 27.29, 28.26, 29.23, 30.16, 31.14, 32.09, 32.99, 33.91, 34.83, 35.68, 36.6, 37.42, 38.3, 39.14, 39.97, 40.83, 41.73, 42.6, 43.46, 44.33, 45.23, 46.06, 46.97, 47.83] 2020-02-01 17:37:58.876: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:37:58.876: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.15 2020-02-01 17:37:58.876: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.11 2020-02-01 17:37:58.877: INFO @evaluate_confidence: Average confidence of all samples 0.46 +- 0.14 2020-02-01 17:37:59.008: INFO @evaluate_confidence: Previous accuracy would be: 42.16 2020-02-01 17:37:59.008: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 17:37:59.027: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [51.9, 52.69, 53.4, 54.13, 54.88, 55.71, 56.48, 57.13, 58.03] 2020-02-01 17:37:59.027: INFO @evaluate_confidence: Dropped ratios are: [47.45, 50.64, 53.56, 56.5, 59.31, 62.05, 64.66, 67.09, 69.48] 2020-02-01 17:37:59.035: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:37:59.035: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.75 +- 0.23 2020-02-01 17:37:59.035: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.16 2020-02-01 17:37:59.035: INFO @evaluate_confidence: Average confidence of all samples 0.72 +- 0.24 2020-02-01 17:37:59.131: INFO @evaluate_confidence: Previous accuracy would be: 87.65 2020-02-01 17:37:59.132: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75] 2020-02-01 17:37:59.139: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.77, 94.04, 94.37, 94.6, 94.77, 94.97, 95.26, 95.45, 95.58, 95.73, 95.85, 95.99, 96.16, 96.31, 96.46, 96.6, 96.77, 96.88, 96.97, 97.14, 97.29, 97.41, 97.57, 97.68, 97.81, 97.91, 98.06, 98.1, 98.18, 98.27, 98.37, 98.45] 2020-02-01 17:37:59.139: INFO @evaluate_confidence: Dropped ratios are: [20.01, 20.92, 21.92, 22.86, 23.74, 24.67, 25.72, 26.55, 27.47, 28.39, 29.3, 30.27, 31.18, 32.06, 33.01, 33.84, 34.71, 35.53, 36.36, 37.21, 38.09, 38.93, 39.72, 40.51, 41.26, 42.08, 42.87, 43.74, 44.63, 45.39, 46.34, 47.25] 2020-02-01 17:37:59.147: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:37:59.147: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.45 +- 0.11 2020-02-01 17:37:59.147: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.09 2020-02-01 17:37:59.147: INFO @evaluate_confidence: Average confidence of all samples 0.43 +- 0.11 2020-02-01 17:37:59.274: INFO @evaluate_confidence: Previous accuracy would be: 46.36 2020-02-01 17:37:59.275: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44] 2020-02-01 17:37:59.276: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [53.38, 53.22, 53.23, 53.42] 2020-02-01 17:37:59.276: INFO @evaluate_confidence: Dropped ratios are: [50.62, 54.88, 59.12, 62.58] 2020-02-01 17:37:59.328: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:38:00.048: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:38:00.128: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:38:00.570: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:38:00.789: INFO @decay_lr : LR updated to `9.13725e-05` 2020-02-01 17:38:00.791: INFO @log_profile : T train: 121.884934 2020-02-01 17:38:00.791: INFO @log_profile : T valid: 5.487331 2020-02-01 17:38:00.791: INFO @log_profile : T read data: 1.934284 2020-02-01 17:38:00.791: INFO @log_profile : T hooks: 7.027530 2020-02-01 17:38:00.791: INFO @main_loop : Epoch 18 done 2020-02-01 17:38:00.791: INFO @main_loop : Training epoch 19 2020-02-01 17:40:11.486: INFO @log_variables: train loss nanmean: 1.072403 2020-02-01 17:40:11.486: INFO @log_variables: train age_loss mean: 7.720238 2020-02-01 17:40:11.486: INFO @log_variables: train gender_loss mean: 0.268597 2020-02-01 17:40:11.486: INFO @log_variables: train age_mae mean: 8.203909 2020-02-01 17:40:11.486: INFO @log_variables: train gender_accuracy mean: 0.881184 2020-02-01 17:40:11.486: INFO @log_variables: train gender_confidence/loss nanmean: 0.066115 2020-02-01 17:40:11.486: INFO @log_variables: train gender_confidence/accuracy mean: 0.780062 2020-02-01 17:40:11.487: INFO @log_variables: train age_confidence/loss mean: 0.060269 2020-02-01 17:40:11.487: INFO @log_variables: train age_confidence/accuracy mean: 0.626907 2020-02-01 17:40:11.487: INFO @log_variables: valid loss nanmean: 0.989208 2020-02-01 17:40:11.487: INFO @log_variables: valid age_loss mean: 6.892644 2020-02-01 17:40:11.487: INFO @log_variables: valid gender_loss mean: 0.260801 2020-02-01 17:40:11.487: INFO @log_variables: valid age_mae mean: 7.376525 2020-02-01 17:40:11.487: INFO @log_variables: valid gender_accuracy mean: 0.884971 2020-02-01 17:40:11.487: INFO @log_variables: valid gender_confidence/loss nanmean: 0.060427 2020-02-01 17:40:11.487: INFO @log_variables: valid gender_confidence/accuracy mean: 0.824140 2020-02-01 17:40:11.487: INFO @log_variables: valid age_confidence/loss mean: 0.065084 2020-02-01 17:40:11.487: INFO @log_variables: valid age_confidence/accuracy mean: 0.563184 2020-02-01 17:40:11.487: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:40:11.495: INFO @metrics_hook: train age_mae: 8.204 +-0.045 (110372) 2020-02-01 17:40:11.502: INFO @metrics_hook: train gender_accuracy: 0.881 +-0.002 (110372) 2020-02-01 17:40:14.263: INFO @metrics_hook: valid age_mae: 7.377 +-0.100 (17639) 2020-02-01 17:40:14.264: INFO @metrics_hook: valid gender_accuracy: 0.885 +-0.005 (17639) 2020-02-01 17:40:17.012: INFO @decay_lr : LR updated to `9.091564e-05` 2020-02-01 17:40:17.309: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:40:17.312: INFO @log_profile : T train: 121.715174 2020-02-01 17:40:17.312: INFO @log_profile : T valid: 5.476871 2020-02-01 17:40:17.312: INFO @log_profile : T read data: 2.807977 2020-02-01 17:40:17.312: INFO @log_profile : T hooks: 6.442874 2020-02-01 17:40:17.312: INFO @main_loop : Epoch 19 done 2020-02-01 17:40:17.312: INFO @main_loop : Training epoch 20 2020-02-01 17:42:29.572: INFO @log_variables: train loss nanmean: 1.057525 2020-02-01 17:42:29.573: INFO @log_variables: train age_loss mean: 7.577818 2020-02-01 17:42:29.573: INFO @log_variables: train gender_loss mean: 0.265254 2020-02-01 17:42:29.573: INFO @log_variables: train age_mae mean: 8.061529 2020-02-01 17:42:29.573: INFO @log_variables: train gender_accuracy mean: 0.885016 2020-02-01 17:42:29.573: INFO @log_variables: train gender_confidence/loss nanmean: 0.066610 2020-02-01 17:42:29.573: INFO @log_variables: train gender_confidence/accuracy mean: 0.781475 2020-02-01 17:42:29.573: INFO @log_variables: train age_confidence/loss mean: 0.060883 2020-02-01 17:42:29.573: INFO @log_variables: train age_confidence/accuracy mean: 0.621009 2020-02-01 17:42:29.573: INFO @log_variables: valid loss nanmean: 0.968894 2020-02-01 17:42:29.573: INFO @log_variables: valid age_loss mean: 6.701613 2020-02-01 17:42:29.573: INFO @log_variables: valid gender_loss mean: 0.257079 2020-02-01 17:42:29.573: INFO @log_variables: valid age_mae mean: 7.183653 2020-02-01 17:42:29.573: INFO @log_variables: valid gender_accuracy mean: 0.885651 2020-02-01 17:42:29.573: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059110 2020-02-01 17:42:29.573: INFO @log_variables: valid gender_confidence/accuracy mean: 0.806452 2020-02-01 17:42:29.573: INFO @log_variables: valid age_confidence/loss mean: 0.066839 2020-02-01 17:42:29.573: INFO @log_variables: valid age_confidence/accuracy mean: 0.553263 2020-02-01 17:42:29.573: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:42:29.580: INFO @metrics_hook: train age_mae: 8.062 +-0.044 (110372) 2020-02-01 17:42:29.588: INFO @metrics_hook: train gender_accuracy: 0.885 +-0.002 (110372) 2020-02-01 17:42:32.383: INFO @metrics_hook: valid age_mae: 7.184 +-0.099 (17639) 2020-02-01 17:42:32.384: INFO @metrics_hook: valid gender_accuracy: 0.886 +-0.005 (17639) 2020-02-01 17:42:33.863: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:42:33.863: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.75 +- 0.23 2020-02-01 17:42:33.863: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.17 2020-02-01 17:42:33.863: INFO @evaluate_confidence: Average confidence of all samples 0.72 +- 0.25 2020-02-01 17:42:34.005: INFO @evaluate_confidence: Previous accuracy would be: 88.50 2020-02-01 17:42:34.006: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74] 2020-02-01 17:42:34.055: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.96, 94.18, 94.39, 94.62, 94.79, 95.01, 95.22, 95.4, 95.55, 95.71, 95.9, 96.06, 96.23, 96.36, 96.48, 96.65, 96.8, 96.93, 97.05, 97.15, 97.28, 97.39, 97.52, 97.63, 97.72, 97.83, 97.92, 98.02, 98.13] 2020-02-01 17:42:34.055: INFO @evaluate_confidence: Dropped ratios are: [21.67, 22.69, 23.69, 24.68, 25.6, 26.58, 27.57, 28.52, 29.42, 30.37, 31.23, 32.1, 33.02, 33.89, 34.76, 35.59, 36.44, 37.29, 38.11, 38.92, 39.77, 40.63, 41.5, 42.32, 43.16, 44.04, 44.91, 45.8, 46.7] 2020-02-01 17:42:34.104: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:42:34.105: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.15 2020-02-01 17:42:34.105: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.11 2020-02-01 17:42:34.105: INFO @evaluate_confidence: Average confidence of all samples 0.46 +- 0.13 2020-02-01 17:42:34.238: INFO @evaluate_confidence: Previous accuracy would be: 43.36 2020-02-01 17:42:34.238: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 17:42:34.255: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [53.0, 53.62, 54.46, 55.16, 55.88, 56.7, 57.46, 58.24] 2020-02-01 17:42:34.255: INFO @evaluate_confidence: Dropped ratios are: [47.22, 50.23, 53.25, 56.13, 58.92, 61.71, 64.31, 66.87] 2020-02-01 17:42:34.263: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:42:34.263: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.22 2020-02-01 17:42:34.263: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.17 2020-02-01 17:42:34.263: INFO @evaluate_confidence: Average confidence of all samples 0.73 +- 0.24 2020-02-01 17:42:34.362: INFO @evaluate_confidence: Previous accuracy would be: 88.57 2020-02-01 17:42:34.363: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 17:42:34.370: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.28, 94.51, 94.68, 94.9, 95.05, 95.25, 95.43, 95.58, 95.8, 95.95, 96.16, 96.27, 96.44, 96.55, 96.71, 96.81, 96.91, 97.02, 97.2, 97.31, 97.43, 97.52, 97.62, 97.71, 97.86, 98.04, 98.13, 98.26, 98.41, 98.53, 98.58] 2020-02-01 17:42:34.370: INFO @evaluate_confidence: Dropped ratios are: [19.62, 20.57, 21.38, 22.25, 23.18, 24.11, 24.92, 25.65, 26.59, 27.41, 28.33, 29.24, 30.0, 30.85, 31.69, 32.54, 33.4, 34.19, 34.97, 35.86, 36.67, 37.47, 38.27, 39.1, 40.0, 41.02, 41.86, 42.79, 43.76, 44.77, 45.68] 2020-02-01 17:42:34.378: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:42:34.378: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.44 +- 0.10 2020-02-01 17:42:34.378: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.08 2020-02-01 17:42:34.378: INFO @evaluate_confidence: Average confidence of all samples 0.42 +- 0.09 2020-02-01 17:42:34.503: INFO @evaluate_confidence: Previous accuracy would be: 47.38 2020-02-01 17:42:34.504: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43] 2020-02-01 17:42:34.505: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [55.82, 56.21, 56.42, 56.28] 2020-02-01 17:42:34.505: INFO @evaluate_confidence: Dropped ratios are: [48.07, 52.18, 56.14, 60.14] 2020-02-01 17:42:34.555: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:42:35.249: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:42:35.332: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:42:35.767: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:42:35.838: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:42:36.528: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:42:36.607: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 17:42:36.609: INFO @evaluate_gender-age_model: groups 0 7.360259 1 7.190516 2 7.259079 3 6.890514 4 7.820219 5 8.590789 6 9.425169 7 11.477484 Name: errors, dtype: float64 2020-02-01 17:42:36.610: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:42:37.059: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:42:37.120: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 17:42:37.121: INFO @evaluate_gender-age_model: groups 0 9.948177 1 8.185922 2 6.707036 3 5.777683 4 7.687664 5 6.216278 6 9.127441 7 14.179413 Name: errors, dtype: float64 2020-02-01 17:42:37.277: INFO @decay_lr : LR updated to `9.046106e-05` 2020-02-01 17:42:37.580: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:42:37.582: INFO @log_profile : T train: 121.497963 2020-02-01 17:42:37.582: INFO @log_profile : T valid: 5.370536 2020-02-01 17:42:37.583: INFO @log_profile : T read data: 2.842676 2020-02-01 17:42:37.583: INFO @log_profile : T hooks: 10.481089 2020-02-01 17:42:37.583: INFO @main_loop : Epoch 20 done 2020-02-01 17:42:37.583: INFO @main_loop : Training epoch 21 2020-02-01 17:44:57.063: INFO @log_variables: train loss nanmean: 1.048227 2020-02-01 17:44:57.063: INFO @log_variables: train age_loss mean: 7.500978 2020-02-01 17:44:57.063: INFO @log_variables: train gender_loss mean: 0.263167 2020-02-01 17:44:57.063: INFO @log_variables: train age_mae mean: 7.984308 2020-02-01 17:44:57.063: INFO @log_variables: train gender_accuracy mean: 0.884476 2020-02-01 17:44:57.063: INFO @log_variables: train gender_confidence/loss nanmean: 0.065960 2020-02-01 17:44:57.063: INFO @log_variables: train gender_confidence/accuracy mean: 0.781404 2020-02-01 17:44:57.063: INFO @log_variables: train age_confidence/loss mean: 0.061118 2020-02-01 17:44:57.063: INFO @log_variables: train age_confidence/accuracy mean: 0.623119 2020-02-01 17:44:57.063: INFO @log_variables: valid loss nanmean: 0.960906 2020-02-01 17:44:57.063: INFO @log_variables: valid age_loss mean: 6.657430 2020-02-01 17:44:57.063: INFO @log_variables: valid gender_loss mean: 0.252928 2020-02-01 17:44:57.063: INFO @log_variables: valid age_mae mean: 7.139322 2020-02-01 17:44:57.063: INFO @log_variables: valid gender_accuracy mean: 0.887125 2020-02-01 17:44:57.063: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059059 2020-02-01 17:44:57.064: INFO @log_variables: valid gender_confidence/accuracy mean: 0.828562 2020-02-01 17:44:57.064: INFO @log_variables: valid age_confidence/loss mean: 0.066691 2020-02-01 17:44:57.064: INFO @log_variables: valid age_confidence/accuracy mean: 0.563467 2020-02-01 17:44:57.064: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:44:57.071: INFO @metrics_hook: train age_mae: 7.984 +-0.044 (110592) 2020-02-01 17:44:57.077: INFO @metrics_hook: train gender_accuracy: 0.884 +-0.002 (110592) 2020-02-01 17:44:59.847: INFO @metrics_hook: valid age_mae: 7.139 +-0.100 (17639) 2020-02-01 17:44:59.848: INFO @metrics_hook: valid gender_accuracy: 0.887 +-0.005 (17639) 2020-02-01 17:45:01.486: INFO @decay_lr : LR updated to `9.0008754e-05` 2020-02-01 17:45:01.787: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:45:01.789: INFO @log_profile : T train: 130.126870 2020-02-01 17:45:01.789: INFO @log_profile : T valid: 6.846169 2020-02-01 17:45:01.789: INFO @log_profile : T read data: 1.818510 2020-02-01 17:45:01.790: INFO @log_profile : T hooks: 5.339463 2020-02-01 17:45:01.790: INFO @main_loop : Epoch 21 done 2020-02-01 17:45:01.790: INFO @main_loop : Training epoch 22 2020-02-01 17:47:21.818: INFO @log_variables: train loss nanmean: 1.034243 2020-02-01 17:47:21.818: INFO @log_variables: train age_loss mean: 7.417203 2020-02-01 17:47:21.818: INFO @log_variables: train gender_loss mean: 0.255980 2020-02-01 17:47:21.818: INFO @log_variables: train age_mae mean: 7.900679 2020-02-01 17:47:21.818: INFO @log_variables: train gender_accuracy mean: 0.888450 2020-02-01 17:47:21.818: INFO @log_variables: train gender_confidence/loss nanmean: 0.065825 2020-02-01 17:47:21.818: INFO @log_variables: train gender_confidence/accuracy mean: 0.785815 2020-02-01 17:47:21.818: INFO @log_variables: train age_confidence/loss mean: 0.061418 2020-02-01 17:47:21.818: INFO @log_variables: train age_confidence/accuracy mean: 0.620637 2020-02-01 17:47:21.818: INFO @log_variables: valid loss nanmean: 0.982401 2020-02-01 17:47:21.818: INFO @log_variables: valid age_loss mean: 6.713374 2020-02-01 17:47:21.818: INFO @log_variables: valid gender_loss mean: 0.270304 2020-02-01 17:47:21.818: INFO @log_variables: valid age_mae mean: 7.193546 2020-02-01 17:47:21.818: INFO @log_variables: valid gender_accuracy mean: 0.877374 2020-02-01 17:47:21.818: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058720 2020-02-01 17:47:21.819: INFO @log_variables: valid gender_confidence/accuracy mean: 0.806395 2020-02-01 17:47:21.819: INFO @log_variables: valid age_confidence/loss mean: 0.067643 2020-02-01 17:47:21.819: INFO @log_variables: valid age_confidence/accuracy mean: 0.552299 2020-02-01 17:47:21.819: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:47:21.826: INFO @metrics_hook: train age_mae: 7.901 +-0.044 (110372) 2020-02-01 17:47:21.832: INFO @metrics_hook: train gender_accuracy: 0.888 +-0.002 (110372) 2020-02-01 17:47:24.602: INFO @metrics_hook: valid age_mae: 7.194 +-0.102 (17639) 2020-02-01 17:47:24.603: INFO @metrics_hook: valid gender_accuracy: 0.877 +-0.005 (17639) 2020-02-01 17:47:26.085: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:47:26.086: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.76 +- 0.24 2020-02-01 17:47:26.086: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.18 2020-02-01 17:47:26.086: INFO @evaluate_confidence: Average confidence of all samples 0.72 +- 0.25 2020-02-01 17:47:26.228: INFO @evaluate_confidence: Previous accuracy would be: 88.84 2020-02-01 17:47:26.228: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 17:47:26.284: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.1, 94.33, 94.53, 94.7, 94.92, 95.13, 95.31, 95.5, 95.68, 95.85, 96.02, 96.16, 96.31, 96.47, 96.6, 96.72, 96.85, 97.0, 97.11, 97.24, 97.33, 97.44, 97.54, 97.64, 97.74, 97.82, 97.91, 98.0, 98.09, 98.18, 98.28, 98.34] 2020-02-01 17:47:26.284: INFO @evaluate_confidence: Dropped ratios are: [20.59, 21.52, 22.44, 23.39, 24.35, 25.3, 26.23, 27.13, 28.03, 28.91, 29.74, 30.55, 31.41, 32.26, 33.1, 33.92, 34.72, 35.57, 36.38, 37.24, 38.07, 38.94, 39.8, 40.62, 41.41, 42.27, 43.11, 43.93, 44.75, 45.63, 46.46, 47.38] 2020-02-01 17:47:26.337: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:47:26.338: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.15 2020-02-01 17:47:26.338: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.12 2020-02-01 17:47:26.338: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.14 2020-02-01 17:47:26.481: INFO @evaluate_confidence: Previous accuracy would be: 44.39 2020-02-01 17:47:26.481: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 17:47:26.497: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [54.08, 54.81, 55.47, 56.2, 56.95, 57.66, 58.36, 59.26] 2020-02-01 17:47:26.498: INFO @evaluate_confidence: Dropped ratios are: [45.88, 48.76, 51.67, 54.46, 57.26, 59.91, 62.58, 65.13] 2020-02-01 17:47:26.505: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:47:26.505: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.23 2020-02-01 17:47:26.505: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.17 2020-02-01 17:47:26.506: INFO @evaluate_confidence: Average confidence of all samples 0.74 +- 0.24 2020-02-01 17:47:26.603: INFO @evaluate_confidence: Previous accuracy would be: 87.74 2020-02-01 17:47:26.603: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 17:47:26.610: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.58, 93.85, 94.05, 94.29, 94.48, 94.72, 94.87, 95.15, 95.38, 95.56, 95.69, 95.84, 96.02, 96.18, 96.36, 96.53, 96.66, 96.79, 96.88, 97.03, 97.14, 97.27, 97.39, 97.49, 97.62, 97.72, 97.87, 98.05, 98.11, 98.18, 98.28] 2020-02-01 17:47:26.610: INFO @evaluate_confidence: Dropped ratios are: [19.77, 20.64, 21.48, 22.35, 23.13, 24.07, 24.8, 25.79, 26.69, 27.54, 28.3, 29.09, 30.0, 30.97, 31.73, 32.52, 33.39, 34.23, 35.06, 35.93, 36.74, 37.56, 38.33, 39.23, 39.99, 40.74, 41.61, 42.5, 43.32, 44.07, 45.05] 2020-02-01 17:47:26.617: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:47:26.618: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.46 +- 0.12 2020-02-01 17:47:26.618: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.10 2020-02-01 17:47:26.618: INFO @evaluate_confidence: Average confidence of all samples 0.43 +- 0.11 2020-02-01 17:47:26.738: INFO @evaluate_confidence: Previous accuracy would be: 48.66 2020-02-01 17:47:26.738: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46] 2020-02-01 17:47:26.740: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.46, 56.74, 56.87, 56.83, 56.81, 56.87] 2020-02-01 17:47:26.740: INFO @evaluate_confidence: Dropped ratios are: [47.93, 51.16, 54.45, 57.7, 60.98, 64.32] 2020-02-01 17:47:26.792: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:47:27.487: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:47:27.565: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:47:27.999: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:47:28.230: INFO @decay_lr : LR updated to `8.955871e-05` 2020-02-01 17:47:28.232: INFO @log_profile : T train: 130.328990 2020-02-01 17:47:28.232: INFO @log_profile : T valid: 6.266482 2020-02-01 17:47:28.232: INFO @log_profile : T read data: 2.779655 2020-02-01 17:47:28.232: INFO @log_profile : T hooks: 6.991556 2020-02-01 17:47:28.232: INFO @main_loop : Epoch 22 done 2020-02-01 17:47:28.232: INFO @main_loop : Training epoch 23 2020-02-01 17:49:45.987: INFO @log_variables: train loss nanmean: 1.024432 2020-02-01 17:49:45.987: INFO @log_variables: train age_loss mean: 7.335687 2020-02-01 17:49:45.987: INFO @log_variables: train gender_loss mean: 0.252831 2020-02-01 17:49:45.987: INFO @log_variables: train age_mae mean: 7.818414 2020-02-01 17:49:45.987: INFO @log_variables: train gender_accuracy mean: 0.890344 2020-02-01 17:49:45.987: INFO @log_variables: train gender_confidence/loss nanmean: 0.065972 2020-02-01 17:49:45.987: INFO @log_variables: train gender_confidence/accuracy mean: 0.786214 2020-02-01 17:49:45.987: INFO @log_variables: train age_confidence/loss mean: 0.061733 2020-02-01 17:49:45.987: INFO @log_variables: train age_confidence/accuracy mean: 0.618644 2020-02-01 17:49:45.987: INFO @log_variables: valid loss nanmean: 0.972977 2020-02-01 17:49:45.987: INFO @log_variables: valid age_loss mean: 6.717570 2020-02-01 17:49:45.987: INFO @log_variables: valid gender_loss mean: 0.259114 2020-02-01 17:49:45.987: INFO @log_variables: valid age_mae mean: 7.200420 2020-02-01 17:49:45.988: INFO @log_variables: valid gender_accuracy mean: 0.884120 2020-02-01 17:49:45.988: INFO @log_variables: valid gender_confidence/loss nanmean: 0.060665 2020-02-01 17:49:45.988: INFO @log_variables: valid gender_confidence/accuracy mean: 0.788990 2020-02-01 17:49:45.988: INFO @log_variables: valid age_confidence/loss mean: 0.066066 2020-02-01 17:49:45.988: INFO @log_variables: valid age_confidence/accuracy mean: 0.574579 2020-02-01 17:49:45.988: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:49:45.995: INFO @metrics_hook: train age_mae: 7.818 +-0.044 (110372) 2020-02-01 17:49:46.002: INFO @metrics_hook: train gender_accuracy: 0.890 +-0.002 (110372) 2020-02-01 17:49:48.708: INFO @metrics_hook: valid age_mae: 7.200 +-0.099 (17639) 2020-02-01 17:49:48.709: INFO @metrics_hook: valid gender_accuracy: 0.884 +-0.005 (17639) 2020-02-01 17:49:50.314: INFO @decay_lr : LR updated to `8.9110916e-05` 2020-02-01 17:49:50.316: INFO @log_profile : T train: 128.769769 2020-02-01 17:49:50.316: INFO @log_profile : T valid: 5.542902 2020-02-01 17:49:50.316: INFO @log_profile : T read data: 2.764321 2020-02-01 17:49:50.316: INFO @log_profile : T hooks: 4.931820 2020-02-01 17:49:50.316: INFO @main_loop : Epoch 23 done 2020-02-01 17:49:50.316: INFO @main_loop : Training epoch 24 2020-02-01 17:52:08.985: INFO @log_variables: train loss nanmean: 1.011112 2020-02-01 17:52:08.985: INFO @log_variables: train age_loss mean: 7.281551 2020-02-01 17:52:08.985: INFO @log_variables: train gender_loss mean: 0.245420 2020-02-01 17:52:08.985: INFO @log_variables: train age_mae mean: 7.764042 2020-02-01 17:52:08.985: INFO @log_variables: train gender_accuracy mean: 0.892940 2020-02-01 17:52:08.985: INFO @log_variables: train gender_confidence/loss nanmean: 0.064500 2020-02-01 17:52:08.985: INFO @log_variables: train gender_confidence/accuracy mean: 0.790654 2020-02-01 17:52:08.985: INFO @log_variables: train age_confidence/loss mean: 0.061545 2020-02-01 17:52:08.985: INFO @log_variables: train age_confidence/accuracy mean: 0.621501 2020-02-01 17:52:08.985: INFO @log_variables: valid loss nanmean: 0.981043 2020-02-01 17:52:08.985: INFO @log_variables: valid age_loss mean: 6.609992 2020-02-01 17:52:08.985: INFO @log_variables: valid gender_loss mean: 0.279429 2020-02-01 17:52:08.985: INFO @log_variables: valid age_mae mean: 7.091041 2020-02-01 17:52:08.985: INFO @log_variables: valid gender_accuracy mean: 0.873179 2020-02-01 17:52:08.985: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058983 2020-02-01 17:52:08.986: INFO @log_variables: valid gender_confidence/accuracy mean: 0.814048 2020-02-01 17:52:08.986: INFO @log_variables: valid age_confidence/loss mean: 0.067125 2020-02-01 17:52:08.986: INFO @log_variables: valid age_confidence/accuracy mean: 0.558195 2020-02-01 17:52:08.986: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:52:08.993: INFO @metrics_hook: train age_mae: 7.764 +-0.043 (110592) 2020-02-01 17:52:09.000: INFO @metrics_hook: train gender_accuracy: 0.893 +-0.002 (110592) 2020-02-01 17:52:11.846: INFO @metrics_hook: valid age_mae: 7.091 +-0.099 (17639) 2020-02-01 17:52:11.848: INFO @metrics_hook: valid gender_accuracy: 0.873 +-0.005 (17639) 2020-02-01 17:52:14.797: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:52:14.797: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.76 +- 0.24 2020-02-01 17:52:14.797: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.18 2020-02-01 17:52:14.797: INFO @evaluate_confidence: Average confidence of all samples 0.73 +- 0.25 2020-02-01 17:52:14.942: INFO @evaluate_confidence: Previous accuracy would be: 89.29 2020-02-01 17:52:14.942: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 17:52:14.996: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.61, 94.79, 94.98, 95.18, 95.35, 95.53, 95.71, 95.87, 96.06, 96.2, 96.34, 96.5, 96.65, 96.78, 96.9, 97.03, 97.13, 97.23, 97.35, 97.47, 97.58, 97.71, 97.8, 97.9, 98.01, 98.1, 98.18, 98.26, 98.39, 98.47, 98.56, 98.63] 2020-02-01 17:52:14.996: INFO @evaluate_confidence: Dropped ratios are: [20.41, 21.32, 22.25, 23.14, 24.01, 24.93, 25.82, 26.65, 27.5, 28.37, 29.23, 30.08, 30.91, 31.69, 32.48, 33.31, 34.06, 34.83, 35.67, 36.49, 37.28, 38.1, 38.88, 39.64, 40.5, 41.34, 42.16, 43.0, 43.85, 44.68, 45.57, 46.38] 2020-02-01 17:52:15.045: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:52:15.046: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.15 2020-02-01 17:52:15.046: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.12 2020-02-01 17:52:15.046: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.14 2020-02-01 17:52:15.185: INFO @evaluate_confidence: Previous accuracy would be: 44.71 2020-02-01 17:52:15.186: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 17:52:15.203: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [54.53, 55.13, 55.87, 56.51, 57.2, 57.94, 58.78, 59.6] 2020-02-01 17:52:15.203: INFO @evaluate_confidence: Dropped ratios are: [45.03, 47.85, 50.73, 53.53, 56.27, 59.04, 61.72, 64.25] 2020-02-01 17:52:15.211: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:52:15.211: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.22 2020-02-01 17:52:15.211: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.18 2020-02-01 17:52:15.211: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.24 2020-02-01 17:52:15.307: INFO @evaluate_confidence: Previous accuracy would be: 87.32 2020-02-01 17:52:15.308: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 17:52:15.315: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.59, 93.76, 94.02, 94.22, 94.45, 94.55, 94.71, 94.84, 95.06, 95.21, 95.36, 95.5, 95.59, 95.77, 95.87, 96.0, 96.21, 96.36, 96.57, 96.69, 96.8, 96.95, 97.06, 97.17, 97.34, 97.49, 97.62, 97.71, 97.81, 97.89, 98.05, 98.22] 2020-02-01 17:52:15.315: INFO @evaluate_confidence: Dropped ratios are: [20.3, 21.08, 21.93, 22.77, 23.45, 24.22, 24.97, 25.81, 26.59, 27.5, 28.3, 29.09, 29.87, 30.6, 31.36, 32.04, 32.9, 33.77, 34.63, 35.48, 36.3, 37.17, 37.98, 38.86, 39.7, 40.56, 41.3, 42.07, 42.89, 43.86, 44.78, 45.75] 2020-02-01 17:52:15.323: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:52:15.323: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.47 +- 0.11 2020-02-01 17:52:15.323: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.10 2020-02-01 17:52:15.323: INFO @evaluate_confidence: Average confidence of all samples 0.45 +- 0.10 2020-02-01 17:52:15.446: INFO @evaluate_confidence: Previous accuracy would be: 48.75 2020-02-01 17:52:15.446: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46] 2020-02-01 17:52:15.448: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.0, 57.34, 57.26, 57.48] 2020-02-01 17:52:15.448: INFO @evaluate_confidence: Dropped ratios are: [46.58, 49.78, 52.87, 56.39] 2020-02-01 17:52:15.506: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:52:16.214: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:52:16.300: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:52:16.751: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:52:16.987: INFO @decay_lr : LR updated to `8.866536e-05` 2020-02-01 17:52:16.988: INFO @log_profile : T train: 130.504049 2020-02-01 17:52:16.988: INFO @log_profile : T valid: 5.646546 2020-02-01 17:52:16.988: INFO @log_profile : T read data: 1.854435 2020-02-01 17:52:16.988: INFO @log_profile : T hooks: 8.591152 2020-02-01 17:52:16.988: INFO @main_loop : Epoch 24 done 2020-02-01 17:52:16.988: INFO @main_loop : Training epoch 25 2020-02-01 17:54:30.585: INFO @log_variables: train loss nanmean: 1.004053 2020-02-01 17:54:30.585: INFO @log_variables: train age_loss mean: 7.189559 2020-02-01 17:54:30.585: INFO @log_variables: train gender_loss mean: 0.245587 2020-02-01 17:54:30.585: INFO @log_variables: train age_mae mean: 7.672061 2020-02-01 17:54:30.585: INFO @log_variables: train gender_accuracy mean: 0.893397 2020-02-01 17:54:30.585: INFO @log_variables: train gender_confidence/loss nanmean: 0.065071 2020-02-01 17:54:30.585: INFO @log_variables: train gender_confidence/accuracy mean: 0.790672 2020-02-01 17:54:30.586: INFO @log_variables: train age_confidence/loss mean: 0.062125 2020-02-01 17:54:30.586: INFO @log_variables: train age_confidence/accuracy mean: 0.616406 2020-02-01 17:54:30.586: INFO @log_variables: valid loss nanmean: 0.934512 2020-02-01 17:54:30.586: INFO @log_variables: valid age_loss mean: 6.463132 2020-02-01 17:54:30.586: INFO @log_variables: valid gender_loss mean: 0.239536 2020-02-01 17:54:30.586: INFO @log_variables: valid age_mae mean: 6.942822 2020-02-01 17:54:30.586: INFO @log_variables: valid gender_accuracy mean: 0.893701 2020-02-01 17:54:30.586: INFO @log_variables: valid gender_confidence/loss nanmean: 0.060294 2020-02-01 17:54:30.586: INFO @log_variables: valid gender_confidence/accuracy mean: 0.818357 2020-02-01 17:54:30.586: INFO @log_variables: valid age_confidence/loss mean: 0.068900 2020-02-01 17:54:30.586: INFO @log_variables: valid age_confidence/accuracy mean: 0.535405 2020-02-01 17:54:30.586: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:54:30.594: INFO @metrics_hook: train age_mae: 7.672 +-0.043 (110372) 2020-02-01 17:54:30.601: INFO @metrics_hook: train gender_accuracy: 0.893 +-0.002 (110372) 2020-02-01 17:54:33.421: INFO @metrics_hook: valid age_mae: 6.943 +-0.101 (17639) 2020-02-01 17:54:33.422: INFO @metrics_hook: valid gender_accuracy: 0.894 +-0.005 (17639) 2020-02-01 17:54:35.039: INFO @decay_lr : LR updated to `8.8222034e-05` 2020-02-01 17:54:35.339: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:54:35.342: INFO @log_profile : T train: 124.450787 2020-02-01 17:54:35.342: INFO @log_profile : T valid: 5.629621 2020-02-01 17:54:35.342: INFO @log_profile : T read data: 2.826180 2020-02-01 17:54:35.342: INFO @log_profile : T hooks: 5.369863 2020-02-01 17:54:35.342: INFO @main_loop : Epoch 25 done 2020-02-01 17:54:35.343: INFO @main_loop : Training epoch 26 2020-02-01 17:56:56.176: INFO @log_variables: train loss nanmean: 0.989683 2020-02-01 17:56:56.176: INFO @log_variables: train age_loss mean: 7.104413 2020-02-01 17:56:56.176: INFO @log_variables: train gender_loss mean: 0.238098 2020-02-01 17:56:56.176: INFO @log_variables: train age_mae mean: 7.586129 2020-02-01 17:56:56.176: INFO @log_variables: train gender_accuracy mean: 0.897302 2020-02-01 17:56:56.176: INFO @log_variables: train gender_confidence/loss nanmean: 0.065038 2020-02-01 17:56:56.176: INFO @log_variables: train gender_confidence/accuracy mean: 0.792837 2020-02-01 17:56:56.176: INFO @log_variables: train age_confidence/loss mean: 0.062336 2020-02-01 17:56:56.176: INFO @log_variables: train age_confidence/accuracy mean: 0.617004 2020-02-01 17:56:56.176: INFO @log_variables: valid loss nanmean: 0.970388 2020-02-01 17:56:56.176: INFO @log_variables: valid age_loss mean: 6.745114 2020-02-01 17:56:56.176: INFO @log_variables: valid gender_loss mean: 0.254783 2020-02-01 17:56:56.176: INFO @log_variables: valid age_mae mean: 7.227248 2020-02-01 17:56:56.177: INFO @log_variables: valid gender_accuracy mean: 0.888542 2020-02-01 17:56:56.177: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059789 2020-02-01 17:56:56.177: INFO @log_variables: valid gender_confidence/accuracy mean: 0.819434 2020-02-01 17:56:56.177: INFO @log_variables: valid age_confidence/loss mean: 0.065786 2020-02-01 17:56:56.177: INFO @log_variables: valid age_confidence/accuracy mean: 0.575146 2020-02-01 17:56:56.177: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:56:56.184: INFO @metrics_hook: train age_mae: 7.586 +-0.042 (110372) 2020-02-01 17:56:56.191: INFO @metrics_hook: train gender_accuracy: 0.897 +-0.002 (110372) 2020-02-01 17:56:58.873: INFO @metrics_hook: valid age_mae: 7.227 +-0.101 (17639) 2020-02-01 17:56:58.875: INFO @metrics_hook: valid gender_accuracy: 0.889 +-0.005 (17639) 2020-02-01 17:57:00.321: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:57:00.321: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.76 +- 0.24 2020-02-01 17:57:00.321: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.18 2020-02-01 17:57:00.322: INFO @evaluate_confidence: Average confidence of all samples 0.73 +- 0.25 2020-02-01 17:57:00.454: INFO @evaluate_confidence: Previous accuracy would be: 89.73 2020-02-01 17:57:00.454: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 17:57:00.505: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.81, 95.01, 95.16, 95.33, 95.48, 95.68, 95.87, 96.04, 96.21, 96.33, 96.47, 96.6, 96.71, 96.84, 96.94, 97.08, 97.18, 97.29, 97.41, 97.53, 97.65, 97.72, 97.82, 97.93, 98.0, 98.07, 98.16, 98.22, 98.3, 98.37, 98.45, 98.52] 2020-02-01 17:57:00.505: INFO @evaluate_confidence: Dropped ratios are: [20.13, 21.0, 21.84, 22.71, 23.58, 24.46, 25.21, 26.04, 26.9, 27.73, 28.54, 29.36, 30.15, 30.93, 31.69, 32.43, 33.24, 34.04, 34.85, 35.63, 36.42, 37.2, 37.96, 38.75, 39.56, 40.41, 41.26, 42.12, 42.99, 43.78, 44.63, 45.54] 2020-02-01 17:57:00.554: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:57:00.554: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.15 2020-02-01 17:57:00.555: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.12 2020-02-01 17:57:00.555: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.14 2020-02-01 17:57:00.688: INFO @evaluate_confidence: Previous accuracy would be: 45.87 2020-02-01 17:57:00.689: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51] 2020-02-01 17:57:00.705: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [55.87, 56.5, 57.18, 57.87, 58.6, 59.25, 60.01, 60.78] 2020-02-01 17:57:00.705: INFO @evaluate_confidence: Dropped ratios are: [45.71, 48.68, 51.53, 54.2, 56.88, 59.58, 62.16, 64.6] 2020-02-01 17:57:00.713: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 17:57:00.713: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.22 2020-02-01 17:57:00.713: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.18 2020-02-01 17:57:00.713: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.24 2020-02-01 17:57:00.806: INFO @evaluate_confidence: Previous accuracy would be: 88.85 2020-02-01 17:57:00.806: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 17:57:00.813: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.22, 94.45, 94.6, 94.77, 94.92, 95.12, 95.29, 95.53, 95.7, 95.84, 96.02, 96.17, 96.41, 96.53, 96.64, 96.79, 96.9, 97.0, 97.1, 97.2, 97.31, 97.41, 97.53, 97.69, 97.74, 97.82, 97.94, 98.01, 98.12, 98.26, 98.41] 2020-02-01 17:57:00.813: INFO @evaluate_confidence: Dropped ratios are: [18.77, 19.56, 20.3, 20.93, 21.7, 22.46, 23.17, 24.08, 24.91, 25.66, 26.5, 27.3, 28.2, 29.04, 29.88, 30.68, 31.56, 32.29, 33.06, 33.98, 34.81, 35.65, 36.53, 37.4, 38.2, 39.07, 40.04, 40.86, 41.77, 42.62, 43.64] 2020-02-01 17:57:00.821: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 17:57:00.821: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.48 +- 0.12 2020-02-01 17:57:00.821: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.10 2020-02-01 17:57:00.821: INFO @evaluate_confidence: Average confidence of all samples 0.46 +- 0.11 2020-02-01 17:57:00.944: INFO @evaluate_confidence: Previous accuracy would be: 47.33 2020-02-01 17:57:00.944: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47] 2020-02-01 17:57:00.946: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [55.88, 56.35, 56.48, 56.68, 57.17] 2020-02-01 17:57:00.946: INFO @evaluate_confidence: Dropped ratios are: [43.98, 47.49, 51.12, 54.79, 58.47] 2020-02-01 17:57:00.997: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 17:57:01.681: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:57:01.762: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 17:57:02.189: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 17:57:02.414: INFO @decay_lr : LR updated to `8.778092e-05` 2020-02-01 17:57:02.415: INFO @log_profile : T train: 130.464210 2020-02-01 17:57:02.415: INFO @log_profile : T valid: 6.890092 2020-02-01 17:57:02.415: INFO @log_profile : T read data: 2.784628 2020-02-01 17:57:02.415: INFO @log_profile : T hooks: 6.857837 2020-02-01 17:57:02.415: INFO @main_loop : Epoch 26 done 2020-02-01 17:57:02.415: INFO @main_loop : Training epoch 27 2020-02-01 17:59:18.857: INFO @log_variables: train loss nanmean: 0.985673 2020-02-01 17:59:18.857: INFO @log_variables: train age_loss mean: 7.071018 2020-02-01 17:59:18.857: INFO @log_variables: train gender_loss mean: 0.237009 2020-02-01 17:59:18.857: INFO @log_variables: train age_mae mean: 7.553214 2020-02-01 17:59:18.857: INFO @log_variables: train gender_accuracy mean: 0.897678 2020-02-01 17:59:18.857: INFO @log_variables: train gender_confidence/loss nanmean: 0.064739 2020-02-01 17:59:18.857: INFO @log_variables: train gender_confidence/accuracy mean: 0.794126 2020-02-01 17:59:18.857: INFO @log_variables: train age_confidence/loss mean: 0.062651 2020-02-01 17:59:18.858: INFO @log_variables: train age_confidence/accuracy mean: 0.615650 2020-02-01 17:59:18.858: INFO @log_variables: valid loss nanmean: 0.917574 2020-02-01 17:59:18.858: INFO @log_variables: valid age_loss mean: 6.383667 2020-02-01 17:59:18.858: INFO @log_variables: valid gender_loss mean: 0.232456 2020-02-01 17:59:18.858: INFO @log_variables: valid age_mae mean: 6.865214 2020-02-01 17:59:18.858: INFO @log_variables: valid gender_accuracy mean: 0.898804 2020-02-01 17:59:18.858: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058873 2020-02-01 17:59:18.858: INFO @log_variables: valid gender_confidence/accuracy mean: 0.833154 2020-02-01 17:59:18.858: INFO @log_variables: valid age_confidence/loss mean: 0.067044 2020-02-01 17:59:18.858: INFO @log_variables: valid age_confidence/accuracy mean: 0.586201 2020-02-01 17:59:18.858: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 17:59:18.866: INFO @metrics_hook: train age_mae: 7.553 +-0.042 (110592) 2020-02-01 17:59:18.873: INFO @metrics_hook: train gender_accuracy: 0.898 +-0.002 (110592) 2020-02-01 17:59:21.636: INFO @metrics_hook: valid age_mae: 6.865 +-0.096 (17639) 2020-02-01 17:59:21.637: INFO @metrics_hook: valid gender_accuracy: 0.899 +-0.005 (17639) 2020-02-01 17:59:23.255: INFO @decay_lr : LR updated to `8.7342014e-05` 2020-02-01 17:59:23.558: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 17:59:23.561: INFO @log_profile : T train: 128.358504 2020-02-01 17:59:23.561: INFO @log_profile : T valid: 5.524719 2020-02-01 17:59:23.561: INFO @log_profile : T read data: 1.880712 2020-02-01 17:59:23.561: INFO @log_profile : T hooks: 5.306979 2020-02-01 17:59:23.561: INFO @main_loop : Epoch 27 done 2020-02-01 17:59:23.561: INFO @main_loop : Training epoch 28 2020-02-01 18:01:35.034: INFO @log_variables: train loss nanmean: 0.972721 2020-02-01 18:01:35.034: INFO @log_variables: train age_loss mean: 6.986535 2020-02-01 18:01:35.034: INFO @log_variables: train gender_loss mean: 0.231364 2020-02-01 18:01:35.034: INFO @log_variables: train age_mae mean: 7.468231 2020-02-01 18:01:35.034: INFO @log_variables: train gender_accuracy mean: 0.900618 2020-02-01 18:01:35.034: INFO @log_variables: train gender_confidence/loss nanmean: 0.064420 2020-02-01 18:01:35.034: INFO @log_variables: train gender_confidence/accuracy mean: 0.798028 2020-02-01 18:01:35.034: INFO @log_variables: train age_confidence/loss mean: 0.062831 2020-02-01 18:01:35.034: INFO @log_variables: train age_confidence/accuracy mean: 0.613453 2020-02-01 18:01:35.035: INFO @log_variables: valid loss nanmean: 0.974739 2020-02-01 18:01:35.035: INFO @log_variables: valid age_loss mean: 6.559104 2020-02-01 18:01:35.035: INFO @log_variables: valid gender_loss mean: 0.273823 2020-02-01 18:01:35.035: INFO @log_variables: valid age_mae mean: 7.041197 2020-02-01 18:01:35.035: INFO @log_variables: valid gender_accuracy mean: 0.887295 2020-02-01 18:01:35.035: INFO @log_variables: valid gender_confidence/loss nanmean: 0.063456 2020-02-01 18:01:35.035: INFO @log_variables: valid gender_confidence/accuracy mean: 0.840297 2020-02-01 18:01:35.035: INFO @log_variables: valid age_confidence/loss mean: 0.066071 2020-02-01 18:01:35.035: INFO @log_variables: valid age_confidence/accuracy mean: 0.574806 2020-02-01 18:01:35.035: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:01:35.042: INFO @metrics_hook: train age_mae: 7.468 +-0.042 (110372) 2020-02-01 18:01:35.049: INFO @metrics_hook: train gender_accuracy: 0.901 +-0.002 (110372) 2020-02-01 18:01:37.788: INFO @metrics_hook: valid age_mae: 7.041 +-0.096 (17639) 2020-02-01 18:01:37.789: INFO @metrics_hook: valid gender_accuracy: 0.887 +-0.005 (17639) 2020-02-01 18:01:39.227: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:01:39.228: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.76 +- 0.24 2020-02-01 18:01:39.228: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.18 2020-02-01 18:01:39.228: INFO @evaluate_confidence: Average confidence of all samples 0.73 +- 0.25 2020-02-01 18:01:39.364: INFO @evaluate_confidence: Previous accuracy would be: 90.06 2020-02-01 18:01:39.364: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75] 2020-02-01 18:01:39.414: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.97, 95.16, 95.32, 95.5, 95.67, 95.82, 95.98, 96.13, 96.29, 96.43, 96.56, 96.68, 96.8, 96.93, 97.04, 97.17, 97.27, 97.37, 97.46, 97.56, 97.65, 97.76, 97.84, 97.93, 98.0, 98.08, 98.17, 98.26, 98.36, 98.45, 98.51, 98.58] 2020-02-01 18:01:39.414: INFO @evaluate_confidence: Dropped ratios are: [19.1, 19.94, 20.76, 21.6, 22.4, 23.22, 24.03, 24.82, 25.63, 26.42, 27.19, 27.98, 28.74, 29.51, 30.28, 31.04, 31.8, 32.54, 33.32, 34.12, 34.89, 35.64, 36.43, 37.22, 38.0, 38.82, 39.6, 40.37, 41.18, 41.97, 42.82, 43.69] 2020-02-01 18:01:39.463: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:01:39.464: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.15 2020-02-01 18:01:39.464: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.12 2020-02-01 18:01:39.464: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.14 2020-02-01 18:01:39.600: INFO @evaluate_confidence: Previous accuracy would be: 46.53 2020-02-01 18:01:39.600: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51] 2020-02-01 18:01:39.616: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.18, 56.85, 57.47, 58.14, 58.86, 59.48, 60.14, 60.94] 2020-02-01 18:01:39.617: INFO @evaluate_confidence: Dropped ratios are: [44.72, 47.54, 50.33, 53.1, 55.83, 58.56, 61.21, 63.75] 2020-02-01 18:01:39.624: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:01:39.624: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.21 2020-02-01 18:01:39.624: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.55 +- 0.20 2020-02-01 18:01:39.625: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.22 2020-02-01 18:01:39.718: INFO @evaluate_confidence: Previous accuracy would be: 88.73 2020-02-01 18:01:39.719: INFO @evaluate_confidence: Possible optimal thresholds are: [0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 18:01:39.725: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [93.8, 93.9, 94.04, 94.21, 94.32, 94.44, 94.56, 94.71, 94.88, 95.05, 95.2, 95.33, 95.44, 95.6, 95.72, 95.87, 96.0, 96.15, 96.28, 96.38, 96.49, 96.66, 96.86, 96.97, 97.07, 97.21, 97.35] 2020-02-01 18:01:39.725: INFO @evaluate_confidence: Dropped ratios are: [18.67, 19.3, 19.96, 20.7, 21.38, 22.12, 22.81, 23.51, 24.2, 24.99, 25.72, 26.61, 27.33, 28.11, 28.87, 29.77, 30.61, 31.46, 32.31, 33.18, 34.03, 34.99, 35.95, 36.83, 37.81, 38.83, 39.85] 2020-02-01 18:01:39.732: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:01:39.733: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.46 +- 0.10 2020-02-01 18:01:39.733: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.09 2020-02-01 18:01:39.733: INFO @evaluate_confidence: Average confidence of all samples 0.44 +- 0.10 2020-02-01 18:01:39.854: INFO @evaluate_confidence: Previous accuracy would be: 47.81 2020-02-01 18:01:39.854: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46] 2020-02-01 18:01:39.855: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.84, 57.39, 57.81, 58.26, 59.02] 2020-02-01 18:01:39.855: INFO @evaluate_confidence: Dropped ratios are: [45.61, 49.32, 52.98, 56.5, 60.24] 2020-02-01 18:01:39.905: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:01:40.579: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:01:40.660: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:01:41.098: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:01:41.326: INFO @decay_lr : LR updated to `8.6905304e-05` 2020-02-01 18:01:41.328: INFO @log_profile : T train: 122.492917 2020-02-01 18:01:41.328: INFO @log_profile : T valid: 5.441563 2020-02-01 18:01:41.328: INFO @log_profile : T read data: 2.841788 2020-02-01 18:01:41.328: INFO @log_profile : T hooks: 6.912833 2020-02-01 18:01:41.328: INFO @main_loop : Epoch 28 done 2020-02-01 18:01:41.328: INFO @main_loop : Training epoch 29 2020-02-01 18:03:52.288: INFO @log_variables: train loss nanmean: 0.968097 2020-02-01 18:03:52.289: INFO @log_variables: train age_loss mean: 6.948171 2020-02-01 18:03:52.289: INFO @log_variables: train gender_loss mean: 0.230119 2020-02-01 18:03:52.289: INFO @log_variables: train age_mae mean: 7.430173 2020-02-01 18:03:52.289: INFO @log_variables: train gender_accuracy mean: 0.900962 2020-02-01 18:03:52.289: INFO @log_variables: train gender_confidence/loss nanmean: 0.064214 2020-02-01 18:03:52.289: INFO @log_variables: train gender_confidence/accuracy mean: 0.797059 2020-02-01 18:03:52.289: INFO @log_variables: train age_confidence/loss mean: 0.063032 2020-02-01 18:03:52.289: INFO @log_variables: train age_confidence/accuracy mean: 0.614757 2020-02-01 18:03:52.289: INFO @log_variables: valid loss nanmean: 0.929476 2020-02-01 18:03:52.289: INFO @log_variables: valid age_loss mean: 6.438422 2020-02-01 18:03:52.289: INFO @log_variables: valid gender_loss mean: 0.238446 2020-02-01 18:03:52.289: INFO @log_variables: valid age_mae mean: 6.919866 2020-02-01 18:03:52.289: INFO @log_variables: valid gender_accuracy mean: 0.893588 2020-02-01 18:03:52.289: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059613 2020-02-01 18:03:52.289: INFO @log_variables: valid gender_confidence/accuracy mean: 0.821305 2020-02-01 18:03:52.289: INFO @log_variables: valid age_confidence/loss mean: 0.067784 2020-02-01 18:03:52.289: INFO @log_variables: valid age_confidence/accuracy mean: 0.557741 2020-02-01 18:03:52.289: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:03:52.296: INFO @metrics_hook: train age_mae: 7.430 +-0.042 (110372) 2020-02-01 18:03:52.303: INFO @metrics_hook: train gender_accuracy: 0.901 +-0.002 (110372) 2020-02-01 18:03:55.017: INFO @metrics_hook: valid age_mae: 6.920 +-0.096 (17639) 2020-02-01 18:03:55.019: INFO @metrics_hook: valid gender_accuracy: 0.894 +-0.005 (17639) 2020-02-01 18:03:56.635: INFO @decay_lr : LR updated to `8.6470776e-05` 2020-02-01 18:03:56.636: INFO @log_profile : T train: 121.801672 2020-02-01 18:03:56.636: INFO @log_profile : T valid: 5.522086 2020-02-01 18:03:56.636: INFO @log_profile : T read data: 2.927572 2020-02-01 18:03:56.636: INFO @log_profile : T hooks: 4.979281 2020-02-01 18:03:56.636: INFO @main_loop : Epoch 29 done 2020-02-01 18:03:56.636: INFO @main_loop : Training epoch 30 2020-02-01 18:06:17.896: INFO @log_variables: train loss nanmean: 0.954972 2020-02-01 18:06:17.896: INFO @log_variables: train age_loss mean: 6.861835 2020-02-01 18:06:17.896: INFO @log_variables: train gender_loss mean: 0.224008 2020-02-01 18:06:17.896: INFO @log_variables: train age_mae mean: 7.343754 2020-02-01 18:06:17.896: INFO @log_variables: train gender_accuracy mean: 0.905328 2020-02-01 18:06:17.896: INFO @log_variables: train gender_confidence/loss nanmean: 0.064332 2020-02-01 18:06:17.896: INFO @log_variables: train gender_confidence/accuracy mean: 0.798801 2020-02-01 18:06:17.896: INFO @log_variables: train age_confidence/loss mean: 0.063193 2020-02-01 18:06:17.896: INFO @log_variables: train age_confidence/accuracy mean: 0.617468 2020-02-01 18:06:17.896: INFO @log_variables: valid loss nanmean: 0.967104 2020-02-01 18:06:17.896: INFO @log_variables: valid age_loss mean: 6.713356 2020-02-01 18:06:17.896: INFO @log_variables: valid gender_loss mean: 0.254205 2020-02-01 18:06:17.896: INFO @log_variables: valid age_mae mean: 7.196815 2020-02-01 18:06:17.896: INFO @log_variables: valid gender_accuracy mean: 0.891661 2020-02-01 18:06:17.896: INFO @log_variables: valid gender_confidence/loss nanmean: 0.060155 2020-02-01 18:06:17.897: INFO @log_variables: valid gender_confidence/accuracy mean: 0.813482 2020-02-01 18:06:17.897: INFO @log_variables: valid age_confidence/loss mean: 0.065548 2020-02-01 18:06:17.897: INFO @log_variables: valid age_confidence/accuracy mean: 0.566472 2020-02-01 18:06:17.897: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:06:17.904: INFO @metrics_hook: train age_mae: 7.344 +-0.041 (110592) 2020-02-01 18:06:17.911: INFO @metrics_hook: train gender_accuracy: 0.905 +-0.002 (110592) 2020-02-01 18:06:20.627: INFO @metrics_hook: valid age_mae: 7.197 +-0.097 (17639) 2020-02-01 18:06:20.628: INFO @metrics_hook: valid gender_accuracy: 0.892 +-0.005 (17639) 2020-02-01 18:06:22.069: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:06:22.069: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.24 2020-02-01 18:06:22.069: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.18 2020-02-01 18:06:22.070: INFO @evaluate_confidence: Average confidence of all samples 0.73 +- 0.25 2020-02-01 18:06:22.206: INFO @evaluate_confidence: Previous accuracy would be: 90.53 2020-02-01 18:06:22.206: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 18:06:22.259: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.26, 95.42, 95.57, 95.72, 95.87, 96.01, 96.15, 96.3, 96.43, 96.57, 96.71, 96.84, 96.98, 97.09, 97.2, 97.32, 97.39, 97.51, 97.62, 97.74, 97.81, 97.9, 97.98, 98.06, 98.14, 98.21, 98.3, 98.4, 98.45, 98.53, 98.58, 98.64, 98.7] 2020-02-01 18:06:22.259: INFO @evaluate_confidence: Dropped ratios are: [18.75, 19.58, 20.43, 21.28, 22.11, 22.91, 23.71, 24.54, 25.35, 26.16, 26.92, 27.7, 28.46, 29.23, 29.96, 30.71, 31.45, 32.24, 32.96, 33.75, 34.52, 35.28, 36.07, 36.85, 37.62, 38.39, 39.17, 39.98, 40.73, 41.57, 42.38, 43.27, 44.13] 2020-02-01 18:06:22.307: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:06:22.307: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.15 2020-02-01 18:06:22.308: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.12 2020-02-01 18:06:22.308: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.14 2020-02-01 18:06:24.159: INFO @evaluate_confidence: Previous accuracy would be: 47.20 2020-02-01 18:06:24.160: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52] 2020-02-01 18:06:24.176: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.4, 58.11, 58.72, 59.5, 60.27, 60.98, 61.63, 62.37] 2020-02-01 18:06:24.176: INFO @evaluate_confidence: Dropped ratios are: [45.91, 48.7, 51.39, 54.09, 56.68, 59.25, 61.76, 64.29] 2020-02-01 18:06:24.183: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:06:24.184: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.22 2020-02-01 18:06:24.184: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.18 2020-02-01 18:06:24.184: INFO @evaluate_confidence: Average confidence of all samples 0.74 +- 0.24 2020-02-01 18:06:24.275: INFO @evaluate_confidence: Previous accuracy would be: 89.17 2020-02-01 18:06:24.275: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 18:06:24.282: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.66, 94.81, 94.98, 95.22, 95.4, 95.56, 95.73, 95.88, 95.98, 96.04, 96.18, 96.28, 96.4, 96.49, 96.61, 96.68, 96.77, 96.95, 97.05, 97.17, 97.26, 97.45, 97.54, 97.67, 97.76, 97.86, 97.96, 98.13, 98.24, 98.33] 2020-02-01 18:06:24.282: INFO @evaluate_confidence: Dropped ratios are: [19.65, 20.51, 21.19, 22.03, 22.93, 23.7, 24.63, 25.61, 26.36, 27.05, 27.91, 28.71, 29.47, 30.29, 31.09, 32.05, 33.03, 33.83, 34.55, 35.38, 36.29, 37.14, 37.99, 38.95, 39.92, 40.76, 41.66, 42.66, 43.66, 44.65] 2020-02-01 18:06:24.290: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:06:24.290: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.45 +- 0.11 2020-02-01 18:06:24.290: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.09 2020-02-01 18:06:24.290: INFO @evaluate_confidence: Average confidence of all samples 0.43 +- 0.11 2020-02-01 18:06:24.415: INFO @evaluate_confidence: Previous accuracy would be: 46.10 2020-02-01 18:06:24.416: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44] 2020-02-01 18:06:24.417: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [54.05, 54.88, 55.42, 56.13] 2020-02-01 18:06:24.417: INFO @evaluate_confidence: Dropped ratios are: [49.49, 53.41, 57.08, 60.76] 2020-02-01 18:06:24.468: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:06:25.149: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:06:25.229: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:06:25.660: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:06:25.731: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:06:26.416: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:06:26.494: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:06:26.495: INFO @evaluate_gender-age_model: groups 0 6.182836 1 6.268554 2 6.725727 3 6.488348 4 7.553465 5 7.818669 6 8.379787 7 10.298033 Name: errors, dtype: float64 2020-02-01 18:06:26.496: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:06:26.934: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:06:26.991: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:06:26.992: INFO @evaluate_gender-age_model: groups 0 8.560225 1 7.561252 2 6.374139 3 5.451225 4 7.056444 5 6.531110 6 10.399293 7 15.448169 Name: errors, dtype: float64 2020-02-01 18:06:27.149: INFO @decay_lr : LR updated to `8.6038424e-05` 2020-02-01 18:06:27.150: INFO @log_profile : T train: 130.875904 2020-02-01 18:06:27.150: INFO @log_profile : T valid: 6.099759 2020-02-01 18:06:27.150: INFO @log_profile : T read data: 1.851288 2020-02-01 18:06:27.150: INFO @log_profile : T hooks: 11.613205 2020-02-01 18:06:27.150: INFO @main_loop : Epoch 30 done 2020-02-01 18:06:27.150: INFO @main_loop : Training epoch 31 2020-02-01 18:08:39.947: INFO @log_variables: train loss nanmean: 0.949709 2020-02-01 18:08:39.947: INFO @log_variables: train age_loss mean: 6.835126 2020-02-01 18:08:39.947: INFO @log_variables: train gender_loss mean: 0.221331 2020-02-01 18:08:39.947: INFO @log_variables: train age_mae mean: 7.317066 2020-02-01 18:08:39.947: INFO @log_variables: train gender_accuracy mean: 0.905982 2020-02-01 18:08:39.948: INFO @log_variables: train gender_confidence/loss nanmean: 0.063667 2020-02-01 18:08:39.948: INFO @log_variables: train gender_confidence/accuracy mean: 0.801807 2020-02-01 18:08:39.948: INFO @log_variables: train age_confidence/loss mean: 0.063457 2020-02-01 18:08:39.948: INFO @log_variables: train age_confidence/accuracy mean: 0.612447 2020-02-01 18:08:39.948: INFO @log_variables: valid loss nanmean: 0.925069 2020-02-01 18:08:39.948: INFO @log_variables: valid age_loss mean: 6.433497 2020-02-01 18:08:39.948: INFO @log_variables: valid gender_loss mean: 0.237072 2020-02-01 18:08:39.948: INFO @log_variables: valid age_mae mean: 6.916326 2020-02-01 18:08:39.948: INFO @log_variables: valid gender_accuracy mean: 0.895459 2020-02-01 18:08:39.948: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057546 2020-02-01 18:08:39.948: INFO @log_variables: valid gender_confidence/accuracy mean: 0.837746 2020-02-01 18:08:39.948: INFO @log_variables: valid age_confidence/loss mean: 0.067139 2020-02-01 18:08:39.948: INFO @log_variables: valid age_confidence/accuracy mean: 0.564261 2020-02-01 18:08:39.948: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:08:39.955: INFO @metrics_hook: train age_mae: 7.317 +-0.041 (110372) 2020-02-01 18:08:39.962: INFO @metrics_hook: train gender_accuracy: 0.906 +-0.002 (110372) 2020-02-01 18:08:42.711: INFO @metrics_hook: valid age_mae: 6.916 +-0.095 (17639) 2020-02-01 18:08:42.712: INFO @metrics_hook: valid gender_accuracy: 0.895 +-0.005 (17639) 2020-02-01 18:08:44.344: INFO @decay_lr : LR updated to `8.560823e-05` 2020-02-01 18:08:44.345: INFO @log_profile : T train: 123.804525 2020-02-01 18:08:44.345: INFO @log_profile : T valid: 5.409627 2020-02-01 18:08:44.345: INFO @log_profile : T read data: 2.881216 2020-02-01 18:08:44.345: INFO @log_profile : T hooks: 5.022821 2020-02-01 18:08:44.345: INFO @main_loop : Epoch 31 done 2020-02-01 18:08:44.346: INFO @main_loop : Training epoch 32 2020-02-01 18:11:02.864: INFO @log_variables: train loss nanmean: 0.945422 2020-02-01 18:11:02.864: INFO @log_variables: train age_loss mean: 6.775348 2020-02-01 18:11:02.864: INFO @log_variables: train gender_loss mean: 0.221758 2020-02-01 18:11:02.864: INFO @log_variables: train age_mae mean: 7.256761 2020-02-01 18:11:02.864: INFO @log_variables: train gender_accuracy mean: 0.906426 2020-02-01 18:11:02.864: INFO @log_variables: train gender_confidence/loss nanmean: 0.064145 2020-02-01 18:11:02.864: INFO @log_variables: train gender_confidence/accuracy mean: 0.800502 2020-02-01 18:11:02.864: INFO @log_variables: train age_confidence/loss mean: 0.063737 2020-02-01 18:11:02.865: INFO @log_variables: train age_confidence/accuracy mean: 0.611840 2020-02-01 18:11:02.865: INFO @log_variables: valid loss nanmean: 0.927241 2020-02-01 18:11:02.865: INFO @log_variables: valid age_loss mean: 6.504734 2020-02-01 18:11:02.865: INFO @log_variables: valid gender_loss mean: 0.231963 2020-02-01 18:11:02.865: INFO @log_variables: valid age_mae mean: 6.987697 2020-02-01 18:11:02.865: INFO @log_variables: valid gender_accuracy mean: 0.896536 2020-02-01 18:11:02.865: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057580 2020-02-01 18:11:02.865: INFO @log_variables: valid gender_confidence/accuracy mean: 0.849254 2020-02-01 18:11:02.865: INFO @log_variables: valid age_confidence/loss mean: 0.067446 2020-02-01 18:11:02.865: INFO @log_variables: valid age_confidence/accuracy mean: 0.565735 2020-02-01 18:11:02.865: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:11:02.872: INFO @metrics_hook: train age_mae: 7.257 +-0.041 (110372) 2020-02-01 18:11:02.879: INFO @metrics_hook: train gender_accuracy: 0.906 +-0.002 (110372) 2020-02-01 18:11:05.598: INFO @metrics_hook: valid age_mae: 6.988 +-0.099 (17639) 2020-02-01 18:11:05.599: INFO @metrics_hook: valid gender_accuracy: 0.897 +-0.005 (17639) 2020-02-01 18:11:07.088: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:11:07.088: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.24 2020-02-01 18:11:07.088: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.18 2020-02-01 18:11:07.089: INFO @evaluate_confidence: Average confidence of all samples 0.74 +- 0.25 2020-02-01 18:11:07.222: INFO @evaluate_confidence: Previous accuracy would be: 90.64 2020-02-01 18:11:07.222: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 18:11:07.277: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.36, 95.51, 95.67, 95.84, 95.99, 96.11, 96.24, 96.36, 96.5, 96.63, 96.78, 96.9, 97.03, 97.15, 97.26, 97.34, 97.43, 97.52, 97.63, 97.74, 97.85, 97.94, 98.03, 98.1, 98.17, 98.25, 98.31, 98.38, 98.47, 98.55, 98.61, 98.66, 98.73] 2020-02-01 18:11:07.277: INFO @evaluate_confidence: Dropped ratios are: [18.79, 19.59, 20.43, 21.24, 22.02, 22.78, 23.55, 24.36, 25.13, 25.91, 26.68, 27.41, 28.17, 28.94, 29.69, 30.44, 31.22, 31.98, 32.71, 33.46, 34.25, 35.0, 35.73, 36.54, 37.37, 38.17, 38.93, 39.73, 40.59, 41.43, 42.26, 43.12, 43.98] 2020-02-01 18:11:07.329: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:11:07.329: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.15 2020-02-01 18:11:07.329: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.12 2020-02-01 18:11:07.329: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.14 2020-02-01 18:11:07.468: INFO @evaluate_confidence: Previous accuracy would be: 47.85 2020-02-01 18:11:07.468: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52] 2020-02-01 18:11:07.485: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.65, 58.21, 58.85, 59.49, 60.11, 60.9, 61.72, 62.39] 2020-02-01 18:11:07.485: INFO @evaluate_confidence: Dropped ratios are: [44.63, 47.5, 50.37, 53.14, 55.82, 58.57, 61.21, 63.75] 2020-02-01 18:11:07.493: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:11:07.493: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.21 2020-02-01 18:11:07.493: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.18 2020-02-01 18:11:07.493: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.23 2020-02-01 18:11:07.588: INFO @evaluate_confidence: Previous accuracy would be: 89.65 2020-02-01 18:11:07.589: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 18:11:07.595: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.92, 95.04, 95.17, 95.32, 95.49, 95.69, 95.86, 96.06, 96.18, 96.28, 96.35, 96.47, 96.61, 96.75, 96.87, 96.99, 97.12, 97.25, 97.34, 97.44, 97.6, 97.74, 97.89, 97.94, 98.03, 98.16, 98.24, 98.32, 98.41, 98.52, 98.6, 98.71] 2020-02-01 18:11:07.596: INFO @evaluate_confidence: Dropped ratios are: [16.59, 17.2, 17.94, 18.75, 19.39, 20.26, 20.94, 21.66, 22.32, 22.97, 23.57, 24.26, 24.98, 25.72, 26.48, 27.21, 27.93, 28.7, 29.54, 30.36, 31.22, 32.05, 32.96, 33.73, 34.54, 35.39, 36.27, 37.05, 38.04, 39.08, 40.14, 41.12] 2020-02-01 18:11:07.603: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:11:07.603: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.11 2020-02-01 18:11:07.603: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 18:11:07.603: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 18:11:07.726: INFO @evaluate_confidence: Previous accuracy would be: 49.18 2020-02-01 18:11:07.726: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 18:11:07.728: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.43, 56.71, 57.01, 57.19, 57.63] 2020-02-01 18:11:07.728: INFO @evaluate_confidence: Dropped ratios are: [45.28, 48.73, 52.02, 55.67, 59.31] 2020-02-01 18:11:07.781: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:11:08.465: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:11:08.547: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:11:08.996: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:11:09.228: INFO @decay_lr : LR updated to `8.518019e-05` 2020-02-01 18:11:09.229: INFO @log_profile : T train: 129.621366 2020-02-01 18:11:09.229: INFO @log_profile : T valid: 5.434324 2020-02-01 18:11:09.229: INFO @log_profile : T read data: 2.770557 2020-02-01 18:11:09.229: INFO @log_profile : T hooks: 6.980191 2020-02-01 18:11:09.229: INFO @main_loop : Epoch 32 done 2020-02-01 18:11:09.229: INFO @main_loop : Training epoch 33 2020-02-01 18:13:19.260: INFO @log_variables: train loss nanmean: 0.932186 2020-02-01 18:13:19.260: INFO @log_variables: train age_loss mean: 6.675931 2020-02-01 18:13:19.260: INFO @log_variables: train gender_loss mean: 0.217702 2020-02-01 18:13:19.260: INFO @log_variables: train age_mae mean: 7.157166 2020-02-01 18:13:19.260: INFO @log_variables: train gender_accuracy mean: 0.907308 2020-02-01 18:13:19.260: INFO @log_variables: train gender_confidence/loss nanmean: 0.063166 2020-02-01 18:13:19.260: INFO @log_variables: train gender_confidence/accuracy mean: 0.804905 2020-02-01 18:13:19.260: INFO @log_variables: train age_confidence/loss mean: 0.064206 2020-02-01 18:13:19.260: INFO @log_variables: train age_confidence/accuracy mean: 0.608860 2020-02-01 18:13:19.261: INFO @log_variables: valid loss nanmean: 0.909520 2020-02-01 18:13:19.261: INFO @log_variables: valid age_loss mean: 6.323944 2020-02-01 18:13:19.261: INFO @log_variables: valid gender_loss mean: 0.230577 2020-02-01 18:13:19.261: INFO @log_variables: valid age_mae mean: 6.805752 2020-02-01 18:13:19.261: INFO @log_variables: valid gender_accuracy mean: 0.898860 2020-02-01 18:13:19.261: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057259 2020-02-01 18:13:19.261: INFO @log_variables: valid gender_confidence/accuracy mean: 0.828448 2020-02-01 18:13:19.261: INFO @log_variables: valid age_confidence/loss mean: 0.067742 2020-02-01 18:13:19.261: INFO @log_variables: valid age_confidence/accuracy mean: 0.561993 2020-02-01 18:13:19.261: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:13:19.268: INFO @metrics_hook: train age_mae: 7.157 +-0.040 (110592) 2020-02-01 18:13:19.275: INFO @metrics_hook: train gender_accuracy: 0.907 +-0.002 (110592) 2020-02-01 18:13:22.009: INFO @metrics_hook: valid age_mae: 6.806 +-0.095 (17639) 2020-02-01 18:13:22.010: INFO @metrics_hook: valid gender_accuracy: 0.899 +-0.005 (17639) 2020-02-01 18:13:23.620: INFO @decay_lr : LR updated to `8.475429e-05` 2020-02-01 18:13:23.922: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 18:13:23.925: INFO @log_profile : T train: 121.941402 2020-02-01 18:13:23.925: INFO @log_profile : T valid: 5.495525 2020-02-01 18:13:23.925: INFO @log_profile : T read data: 1.892810 2020-02-01 18:13:23.925: INFO @log_profile : T hooks: 5.289571 2020-02-01 18:13:23.926: INFO @main_loop : Epoch 33 done 2020-02-01 18:13:23.926: INFO @main_loop : Training epoch 34 2020-02-01 18:15:34.432: INFO @log_variables: train loss nanmean: 0.928237 2020-02-01 18:15:34.432: INFO @log_variables: train age_loss mean: 6.661848 2020-02-01 18:15:34.433: INFO @log_variables: train gender_loss mean: 0.213793 2020-02-01 18:15:34.433: INFO @log_variables: train age_mae mean: 7.142839 2020-02-01 18:15:34.433: INFO @log_variables: train gender_accuracy mean: 0.910285 2020-02-01 18:15:34.433: INFO @log_variables: train gender_confidence/loss nanmean: 0.064249 2020-02-01 18:15:34.433: INFO @log_variables: train gender_confidence/accuracy mean: 0.802939 2020-02-01 18:15:34.433: INFO @log_variables: train age_confidence/loss mean: 0.064009 2020-02-01 18:15:34.433: INFO @log_variables: train age_confidence/accuracy mean: 0.613661 2020-02-01 18:15:34.433: INFO @log_variables: valid loss nanmean: 0.910761 2020-02-01 18:15:34.433: INFO @log_variables: valid age_loss mean: 6.437362 2020-02-01 18:15:34.433: INFO @log_variables: valid gender_loss mean: 0.221182 2020-02-01 18:15:34.433: INFO @log_variables: valid age_mae mean: 6.919693 2020-02-01 18:15:34.433: INFO @log_variables: valid gender_accuracy mean: 0.903736 2020-02-01 18:15:34.433: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058396 2020-02-01 18:15:34.433: INFO @log_variables: valid gender_confidence/accuracy mean: 0.848347 2020-02-01 18:15:34.433: INFO @log_variables: valid age_confidence/loss mean: 0.066075 2020-02-01 18:15:34.434: INFO @log_variables: valid age_confidence/accuracy mean: 0.577811 2020-02-01 18:15:34.434: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:15:34.441: INFO @metrics_hook: train age_mae: 7.143 +-0.040 (110372) 2020-02-01 18:15:34.448: INFO @metrics_hook: train gender_accuracy: 0.910 +-0.002 (110372) 2020-02-01 18:15:37.227: INFO @metrics_hook: valid age_mae: 6.920 +-0.094 (17639) 2020-02-01 18:15:37.228: INFO @metrics_hook: valid gender_accuracy: 0.904 +-0.004 (17639) 2020-02-01 18:15:38.703: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:15:38.703: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.24 2020-02-01 18:15:38.703: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.19 2020-02-01 18:15:38.704: INFO @evaluate_confidence: Average confidence of all samples 0.74 +- 0.25 2020-02-01 18:15:38.831: INFO @evaluate_confidence: Previous accuracy would be: 91.03 2020-02-01 18:15:38.831: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 18:15:38.883: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.62, 95.79, 95.96, 96.11, 96.25, 96.36, 96.47, 96.61, 96.73, 96.84, 96.96, 97.09, 97.2, 97.29, 97.39, 97.49, 97.58, 97.66, 97.76, 97.84, 97.92, 98.0, 98.06, 98.14, 98.24, 98.3, 98.38, 98.45, 98.49, 98.54, 98.62, 98.67, 98.73] 2020-02-01 18:15:38.883: INFO @evaluate_confidence: Dropped ratios are: [18.6, 19.45, 20.21, 20.98, 21.73, 22.49, 23.26, 23.99, 24.76, 25.5, 26.23, 26.99, 27.73, 28.39, 29.12, 29.86, 30.59, 31.34, 32.12, 32.89, 33.62, 34.37, 35.1, 35.85, 36.62, 37.39, 38.22, 39.03, 39.86, 40.69, 41.54, 42.35, 43.23] 2020-02-01 18:15:38.932: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:15:38.932: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.15 2020-02-01 18:15:38.932: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.12 2020-02-01 18:15:38.933: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.14 2020-02-01 18:15:39.065: INFO @evaluate_confidence: Previous accuracy would be: 48.31 2020-02-01 18:15:39.066: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52] 2020-02-01 18:15:39.082: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.09, 58.69, 59.39, 60.14, 60.78, 61.51, 62.22, 62.98] 2020-02-01 18:15:39.083: INFO @evaluate_confidence: Dropped ratios are: [44.03, 46.89, 49.78, 52.59, 55.43, 57.96, 60.52, 63.06] 2020-02-01 18:15:39.090: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:15:39.090: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.21 2020-02-01 18:15:39.090: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.19 2020-02-01 18:15:39.090: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.23 2020-02-01 18:15:39.184: INFO @evaluate_confidence: Previous accuracy would be: 90.37 2020-02-01 18:15:39.185: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 18:15:39.192: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.12, 95.23, 95.36, 95.49, 95.68, 95.85, 95.96, 96.06, 96.18, 96.32, 96.44, 96.54, 96.67, 96.78, 96.9, 97.04, 97.1, 97.19, 97.25, 97.3, 97.35, 97.46, 97.61, 97.72, 97.81, 97.93, 98.03, 98.09, 98.2, 98.32, 98.4, 98.54, 98.67] 2020-02-01 18:15:39.192: INFO @evaluate_confidence: Dropped ratios are: [15.85, 16.49, 17.09, 17.8, 18.49, 19.2, 19.81, 20.48, 21.28, 21.95, 22.6, 23.37, 24.19, 24.89, 25.65, 26.45, 27.17, 27.97, 28.76, 29.54, 30.31, 31.15, 32.07, 33.0, 33.85, 34.82, 35.8, 36.8, 37.71, 38.67, 39.66, 40.64, 41.77] 2020-02-01 18:15:39.200: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:15:39.200: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.11 2020-02-01 18:15:39.200: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.09 2020-02-01 18:15:39.200: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.10 2020-02-01 18:15:39.323: INFO @evaluate_confidence: Previous accuracy would be: 47.58 2020-02-01 18:15:39.323: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 18:15:39.325: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [55.32, 55.87, 56.4, 57.19, 57.69] 2020-02-01 18:15:39.325: INFO @evaluate_confidence: Dropped ratios are: [47.85, 51.91, 55.96, 60.07, 63.85] 2020-02-01 18:15:39.376: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:15:40.064: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:15:40.140: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:15:40.581: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:15:40.815: INFO @decay_lr : LR updated to `8.433052e-05` 2020-02-01 18:15:41.130: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 18:15:41.133: INFO @log_profile : T train: 121.569049 2020-02-01 18:15:41.133: INFO @log_profile : T valid: 5.412558 2020-02-01 18:15:41.133: INFO @log_profile : T read data: 2.838092 2020-02-01 18:15:41.133: INFO @log_profile : T hooks: 7.312400 2020-02-01 18:15:41.133: INFO @main_loop : Epoch 34 done 2020-02-01 18:15:41.133: INFO @main_loop : Training epoch 35 2020-02-01 18:17:52.014: INFO @log_variables: train loss nanmean: 0.917784 2020-02-01 18:17:52.015: INFO @log_variables: train age_loss mean: 6.602466 2020-02-01 18:17:52.015: INFO @log_variables: train gender_loss mean: 0.209288 2020-02-01 18:17:52.015: INFO @log_variables: train age_mae mean: 7.084010 2020-02-01 18:17:52.015: INFO @log_variables: train gender_accuracy mean: 0.911182 2020-02-01 18:17:52.015: INFO @log_variables: train gender_confidence/loss nanmean: 0.063100 2020-02-01 18:17:52.015: INFO @log_variables: train gender_confidence/accuracy mean: 0.807859 2020-02-01 18:17:52.015: INFO @log_variables: train age_confidence/loss mean: 0.064199 2020-02-01 18:17:52.015: INFO @log_variables: train age_confidence/accuracy mean: 0.609702 2020-02-01 18:17:52.015: INFO @log_variables: valid loss nanmean: 0.879178 2020-02-01 18:17:52.015: INFO @log_variables: valid age_loss mean: 6.178281 2020-02-01 18:17:52.015: INFO @log_variables: valid gender_loss mean: 0.212708 2020-02-01 18:17:52.015: INFO @log_variables: valid age_mae mean: 6.660271 2020-02-01 18:17:52.015: INFO @log_variables: valid gender_accuracy mean: 0.907421 2020-02-01 18:17:52.015: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056658 2020-02-01 18:17:52.015: INFO @log_variables: valid gender_confidence/accuracy mean: 0.851069 2020-02-01 18:17:52.015: INFO @log_variables: valid age_confidence/loss mean: 0.067487 2020-02-01 18:17:52.015: INFO @log_variables: valid age_confidence/accuracy mean: 0.569930 2020-02-01 18:17:52.016: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:17:52.023: INFO @metrics_hook: train age_mae: 7.084 +-0.040 (110372) 2020-02-01 18:17:52.031: INFO @metrics_hook: train gender_accuracy: 0.911 +-0.002 (110372) 2020-02-01 18:17:54.886: INFO @metrics_hook: valid age_mae: 6.660 +-0.091 (17639) 2020-02-01 18:17:54.887: INFO @metrics_hook: valid gender_accuracy: 0.907 +-0.004 (17639) 2020-02-01 18:17:56.538: INFO @decay_lr : LR updated to `8.390887e-05` 2020-02-01 18:17:56.841: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 18:17:56.844: INFO @log_profile : T train: 121.853566 2020-02-01 18:17:56.844: INFO @log_profile : T valid: 5.452929 2020-02-01 18:17:56.844: INFO @log_profile : T read data: 2.895949 2020-02-01 18:17:56.844: INFO @log_profile : T hooks: 5.429587 2020-02-01 18:17:56.844: INFO @main_loop : Epoch 35 done 2020-02-01 18:17:56.844: INFO @main_loop : Training epoch 36 2020-02-01 18:20:07.574: INFO @log_variables: train loss nanmean: 0.915064 2020-02-01 18:20:07.574: INFO @log_variables: train age_loss mean: 6.560252 2020-02-01 18:20:07.574: INFO @log_variables: train gender_loss mean: 0.210521 2020-02-01 18:20:07.574: INFO @log_variables: train age_mae mean: 7.041063 2020-02-01 18:20:07.574: INFO @log_variables: train gender_accuracy mean: 0.911340 2020-02-01 18:20:07.574: INFO @log_variables: train gender_confidence/loss nanmean: 0.062894 2020-02-01 18:20:07.574: INFO @log_variables: train gender_confidence/accuracy mean: 0.806485 2020-02-01 18:20:07.574: INFO @log_variables: train age_confidence/loss mean: 0.064401 2020-02-01 18:20:07.574: INFO @log_variables: train age_confidence/accuracy mean: 0.610637 2020-02-01 18:20:07.574: INFO @log_variables: valid loss nanmean: 0.887360 2020-02-01 18:20:07.575: INFO @log_variables: valid age_loss mean: 6.244063 2020-02-01 18:20:07.575: INFO @log_variables: valid gender_loss mean: 0.214080 2020-02-01 18:20:07.575: INFO @log_variables: valid age_mae mean: 6.724600 2020-02-01 18:20:07.575: INFO @log_variables: valid gender_accuracy mean: 0.908385 2020-02-01 18:20:07.575: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057211 2020-02-01 18:20:07.575: INFO @log_variables: valid gender_confidence/accuracy mean: 0.853676 2020-02-01 18:20:07.575: INFO @log_variables: valid age_confidence/loss mean: 0.067889 2020-02-01 18:20:07.575: INFO @log_variables: valid age_confidence/accuracy mean: 0.551959 2020-02-01 18:20:07.575: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:20:07.582: INFO @metrics_hook: train age_mae: 7.041 +-0.040 (110591) 2020-02-01 18:20:07.589: INFO @metrics_hook: train gender_accuracy: 0.911 +-0.002 (110591) 2020-02-01 18:20:10.342: INFO @metrics_hook: valid age_mae: 6.725 +-0.094 (17639) 2020-02-01 18:20:10.343: INFO @metrics_hook: valid gender_accuracy: 0.908 +-0.004 (17639) 2020-02-01 18:20:11.794: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:20:11.795: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.77 +- 0.24 2020-02-01 18:20:11.795: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.19 2020-02-01 18:20:11.795: INFO @evaluate_confidence: Average confidence of all samples 0.74 +- 0.25 2020-02-01 18:20:11.923: INFO @evaluate_confidence: Previous accuracy would be: 91.13 2020-02-01 18:20:11.923: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76] 2020-02-01 18:20:11.979: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.56, 95.72, 95.87, 96.06, 96.22, 96.34, 96.48, 96.6, 96.73, 96.87, 96.98, 97.06, 97.17, 97.27, 97.37, 97.49, 97.6, 97.7, 97.78, 97.89, 97.95, 98.04, 98.08, 98.16, 98.21, 98.3, 98.36, 98.42, 98.48, 98.53, 98.61, 98.67, 98.73, 98.79] 2020-02-01 18:20:11.979: INFO @evaluate_confidence: Dropped ratios are: [17.71, 18.48, 19.23, 19.99, 20.78, 21.51, 22.26, 22.98, 23.74, 24.45, 25.15, 25.85, 26.5, 27.23, 27.93, 28.65, 29.35, 30.05, 30.75, 31.48, 32.15, 32.9, 33.63, 34.38, 35.1, 35.87, 36.62, 37.35, 38.07, 38.92, 39.67, 40.54, 41.34, 42.2] 2020-02-01 18:20:12.032: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:20:12.032: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.15 2020-02-01 18:20:12.032: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.12 2020-02-01 18:20:12.032: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.14 2020-02-01 18:20:12.171: INFO @evaluate_confidence: Previous accuracy would be: 48.86 2020-02-01 18:20:12.172: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52] 2020-02-01 18:20:12.188: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.14, 58.79, 59.44, 60.08, 60.85, 61.61, 62.37, 63.13] 2020-02-01 18:20:12.188: INFO @evaluate_confidence: Dropped ratios are: [43.29, 46.14, 48.97, 51.75, 54.48, 57.25, 59.88, 62.43] 2020-02-01 18:20:12.196: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:20:12.196: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.21 2020-02-01 18:20:12.196: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.20 2020-02-01 18:20:12.196: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.23 2020-02-01 18:20:12.294: INFO @evaluate_confidence: Previous accuracy would be: 90.84 2020-02-01 18:20:12.294: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 18:20:12.302: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.53, 95.69, 95.81, 95.99, 96.11, 96.18, 96.33, 96.43, 96.52, 96.66, 96.78, 96.88, 96.95, 97.08, 97.19, 97.31, 97.42, 97.49, 97.62, 97.71, 97.77, 97.82, 97.95, 98.01, 98.07, 98.13, 98.2, 98.28, 98.38, 98.46, 98.56, 98.63, 98.72, 98.83, 98.85] 2020-02-01 18:20:12.302: INFO @evaluate_confidence: Dropped ratios are: [15.53, 16.23, 16.83, 17.52, 18.09, 18.58, 19.25, 19.89, 20.52, 21.25, 21.85, 22.4, 22.98, 23.61, 24.27, 24.92, 25.59, 26.28, 27.0, 27.71, 28.34, 29.12, 29.84, 30.64, 31.42, 32.2, 32.92, 33.65, 34.53, 35.39, 36.3, 37.13, 38.15, 39.08, 40.05] 2020-02-01 18:20:12.310: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:20:12.310: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.46 +- 0.11 2020-02-01 18:20:12.310: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.09 2020-02-01 18:20:12.310: INFO @evaluate_confidence: Average confidence of all samples 0.44 +- 0.10 2020-02-01 18:20:12.435: INFO @evaluate_confidence: Previous accuracy would be: 49.82 2020-02-01 18:20:12.435: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46] 2020-02-01 18:20:12.437: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.27, 58.97, 59.55, 60.02, 60.55] 2020-02-01 18:20:12.437: INFO @evaluate_confidence: Dropped ratios are: [46.11, 50.62, 55.46, 59.7, 63.77] 2020-02-01 18:20:12.486: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:20:13.175: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:20:13.255: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:20:15.845: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:20:16.082: INFO @decay_lr : LR updated to `8.348932e-05` 2020-02-01 18:20:16.401: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 18:20:16.404: INFO @log_profile : T train: 122.703479 2020-02-01 18:20:16.404: INFO @log_profile : T valid: 5.484451 2020-02-01 18:20:16.404: INFO @log_profile : T read data: 1.856210 2020-02-01 18:20:16.404: INFO @log_profile : T hooks: 9.438750 2020-02-01 18:20:16.404: INFO @main_loop : Epoch 36 done 2020-02-01 18:20:16.404: INFO @main_loop : Training epoch 37 2020-02-01 18:22:37.416: INFO @log_variables: train loss nanmean: 0.905602 2020-02-01 18:22:37.416: INFO @log_variables: train age_loss mean: 6.484393 2020-02-01 18:22:37.417: INFO @log_variables: train gender_loss mean: 0.207275 2020-02-01 18:22:37.417: INFO @log_variables: train age_mae mean: 6.965168 2020-02-01 18:22:37.417: INFO @log_variables: train gender_accuracy mean: 0.912161 2020-02-01 18:22:37.417: INFO @log_variables: train gender_confidence/loss nanmean: 0.063055 2020-02-01 18:22:37.417: INFO @log_variables: train gender_confidence/accuracy mean: 0.807043 2020-02-01 18:22:37.417: INFO @log_variables: train age_confidence/loss mean: 0.064625 2020-02-01 18:22:37.417: INFO @log_variables: train age_confidence/accuracy mean: 0.609530 2020-02-01 18:22:37.417: INFO @log_variables: valid loss nanmean: 0.905553 2020-02-01 18:22:37.417: INFO @log_variables: valid age_loss mean: 6.481867 2020-02-01 18:22:37.417: INFO @log_variables: valid gender_loss mean: 0.212713 2020-02-01 18:22:37.417: INFO @log_variables: valid age_mae mean: 6.962858 2020-02-01 18:22:37.417: INFO @log_variables: valid gender_accuracy mean: 0.908271 2020-02-01 18:22:37.417: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054993 2020-02-01 18:22:37.417: INFO @log_variables: valid gender_confidence/accuracy mean: 0.842451 2020-02-01 18:22:37.417: INFO @log_variables: valid age_confidence/loss mean: 0.067923 2020-02-01 18:22:37.417: INFO @log_variables: valid age_confidence/accuracy mean: 0.542888 2020-02-01 18:22:37.417: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:22:37.429: INFO @metrics_hook: train age_mae: 6.965 +-0.039 (110372) 2020-02-01 18:22:37.440: INFO @metrics_hook: train gender_accuracy: 0.912 +-0.002 (110372) 2020-02-01 18:22:40.152: INFO @metrics_hook: valid age_mae: 6.963 +-0.100 (17639) 2020-02-01 18:22:40.154: INFO @metrics_hook: valid gender_accuracy: 0.908 +-0.004 (17639) 2020-02-01 18:22:41.864: INFO @decay_lr : LR updated to `8.307188e-05` 2020-02-01 18:22:41.866: INFO @log_profile : T train: 130.693241 2020-02-01 18:22:41.866: INFO @log_profile : T valid: 6.776514 2020-02-01 18:22:41.866: INFO @log_profile : T read data: 2.851322 2020-02-01 18:22:41.866: INFO @log_profile : T hooks: 5.064289 2020-02-01 18:22:41.866: INFO @main_loop : Epoch 37 done 2020-02-01 18:22:41.866: INFO @main_loop : Training epoch 38 2020-02-01 18:24:58.861: INFO @log_variables: train loss nanmean: 0.902444 2020-02-01 18:24:58.861: INFO @log_variables: train age_loss mean: 6.489679 2020-02-01 18:24:58.861: INFO @log_variables: train gender_loss mean: 0.203883 2020-02-01 18:24:58.861: INFO @log_variables: train age_mae mean: 6.970382 2020-02-01 18:24:58.861: INFO @log_variables: train gender_accuracy mean: 0.914723 2020-02-01 18:24:58.861: INFO @log_variables: train gender_confidence/loss nanmean: 0.062651 2020-02-01 18:24:58.861: INFO @log_variables: train gender_confidence/accuracy mean: 0.810276 2020-02-01 18:24:58.861: INFO @log_variables: train age_confidence/loss mean: 0.064474 2020-02-01 18:24:58.861: INFO @log_variables: train age_confidence/accuracy mean: 0.609547 2020-02-01 18:24:58.861: INFO @log_variables: valid loss nanmean: 0.897991 2020-02-01 18:24:58.861: INFO @log_variables: valid age_loss mean: 6.292094 2020-02-01 18:24:58.861: INFO @log_variables: valid gender_loss mean: 0.221244 2020-02-01 18:24:58.861: INFO @log_variables: valid age_mae mean: 6.772944 2020-02-01 18:24:58.861: INFO @log_variables: valid gender_accuracy mean: 0.904983 2020-02-01 18:24:58.861: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056225 2020-02-01 18:24:58.861: INFO @log_variables: valid gender_confidence/accuracy mean: 0.840751 2020-02-01 18:24:58.861: INFO @log_variables: valid age_confidence/loss mean: 0.068626 2020-02-01 18:24:58.862: INFO @log_variables: valid age_confidence/accuracy mean: 0.560066 2020-02-01 18:24:58.862: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:24:58.869: INFO @metrics_hook: train age_mae: 6.970 +-0.039 (110592) 2020-02-01 18:24:58.875: INFO @metrics_hook: train gender_accuracy: 0.915 +-0.002 (110592) 2020-02-01 18:25:01.585: INFO @metrics_hook: valid age_mae: 6.773 +-0.097 (17639) 2020-02-01 18:25:01.586: INFO @metrics_hook: valid gender_accuracy: 0.905 +-0.004 (17639) 2020-02-01 18:25:03.068: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:25:03.068: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:25:03.068: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.19 2020-02-01 18:25:03.069: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.25 2020-02-01 18:25:03.198: INFO @evaluate_confidence: Previous accuracy would be: 91.47 2020-02-01 18:25:03.198: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:25:03.253: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.88, 96.02, 96.16, 96.27, 96.4, 96.53, 96.67, 96.78, 96.91, 97.01, 97.1, 97.21, 97.29, 97.39, 97.48, 97.59, 97.69, 97.76, 97.87, 97.95, 98.02, 98.09, 98.16, 98.25, 98.31, 98.38, 98.45, 98.51, 98.58, 98.63, 98.7, 98.75, 98.8, 98.86, 98.9] 2020-02-01 18:25:03.253: INFO @evaluate_confidence: Dropped ratios are: [17.47, 18.2, 18.98, 19.66, 20.33, 21.07, 21.83, 22.51, 23.22, 23.91, 24.63, 25.31, 25.97, 26.65, 27.34, 28.03, 28.73, 29.38, 30.17, 30.88, 31.56, 32.3, 32.98, 33.69, 34.42, 35.12, 35.83, 36.59, 37.35, 38.13, 38.9, 39.69, 40.51, 41.41, 42.27] 2020-02-01 18:25:03.302: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:25:03.302: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.15 2020-02-01 18:25:03.303: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.12 2020-02-01 18:25:03.303: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.14 2020-02-01 18:25:03.442: INFO @evaluate_confidence: Previous accuracy would be: 49.05 2020-02-01 18:25:03.442: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52] 2020-02-01 18:25:03.459: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.22, 58.96, 59.66, 60.28, 61.03, 61.66, 62.34, 63.1] 2020-02-01 18:25:03.459: INFO @evaluate_confidence: Dropped ratios are: [42.9, 45.84, 48.68, 51.53, 54.37, 57.09, 59.72, 62.39] 2020-02-01 18:25:03.467: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:25:03.467: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.22 2020-02-01 18:25:03.467: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.19 2020-02-01 18:25:03.467: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.24 2020-02-01 18:25:03.563: INFO @evaluate_confidence: Previous accuracy would be: 90.50 2020-02-01 18:25:03.563: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 18:25:03.571: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.35, 95.55, 95.71, 95.82, 95.94, 96.02, 96.13, 96.23, 96.37, 96.51, 96.59, 96.69, 96.77, 96.93, 97.03, 97.11, 97.25, 97.32, 97.4, 97.51, 97.56, 97.68, 97.83, 97.94, 98.0, 98.1, 98.17, 98.31, 98.38, 98.47, 98.53, 98.6, 98.69, 98.76, 98.82, 98.9, 98.98] 2020-02-01 18:25:03.571: INFO @evaluate_confidence: Dropped ratios are: [15.51, 16.08, 16.7, 17.35, 17.99, 18.65, 19.18, 19.8, 20.35, 20.99, 21.63, 22.31, 23.09, 23.67, 24.3, 24.94, 25.61, 26.21, 27.11, 27.97, 28.79, 29.49, 30.3, 31.06, 31.79, 32.48, 33.13, 33.94, 34.84, 35.64, 36.41, 37.3, 38.14, 39.03, 39.96, 40.99, 42.08] 2020-02-01 18:25:03.579: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:25:03.579: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.12 2020-02-01 18:25:03.579: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.10 2020-02-01 18:25:03.579: INFO @evaluate_confidence: Average confidence of all samples 0.46 +- 0.12 2020-02-01 18:25:03.709: INFO @evaluate_confidence: Previous accuracy would be: 50.99 2020-02-01 18:25:03.709: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 18:25:03.711: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.28, 59.84, 60.29, 60.15, 60.41, 60.62] 2020-02-01 18:25:03.711: INFO @evaluate_confidence: Dropped ratios are: [46.69, 50.16, 53.91, 57.45, 60.6, 63.92] 2020-02-01 18:25:03.763: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:25:04.458: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:25:04.545: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:25:05.003: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:25:05.239: INFO @decay_lr : LR updated to `8.2656516e-05` 2020-02-01 18:25:05.241: INFO @log_profile : T train: 129.029074 2020-02-01 18:25:05.241: INFO @log_profile : T valid: 5.437914 2020-02-01 18:25:05.241: INFO @log_profile : T read data: 1.845584 2020-02-01 18:25:05.241: INFO @log_profile : T hooks: 6.986341 2020-02-01 18:25:05.241: INFO @main_loop : Epoch 38 done 2020-02-01 18:25:05.241: INFO @main_loop : Training epoch 39 2020-02-01 18:27:22.628: INFO @log_variables: train loss nanmean: 0.899797 2020-02-01 18:27:22.628: INFO @log_variables: train age_loss mean: 6.467361 2020-02-01 18:27:22.628: INFO @log_variables: train gender_loss mean: 0.203530 2020-02-01 18:27:22.628: INFO @log_variables: train age_mae mean: 6.947696 2020-02-01 18:27:22.628: INFO @log_variables: train gender_accuracy mean: 0.914371 2020-02-01 18:27:22.628: INFO @log_variables: train gender_confidence/loss nanmean: 0.062199 2020-02-01 18:27:22.628: INFO @log_variables: train gender_confidence/accuracy mean: 0.810350 2020-02-01 18:27:22.628: INFO @log_variables: train age_confidence/loss mean: 0.064629 2020-02-01 18:27:22.628: INFO @log_variables: train age_confidence/accuracy mean: 0.612057 2020-02-01 18:27:22.629: INFO @log_variables: valid loss nanmean: 0.886404 2020-02-01 18:27:22.629: INFO @log_variables: valid age_loss mean: 6.243788 2020-02-01 18:27:22.629: INFO @log_variables: valid gender_loss mean: 0.214658 2020-02-01 18:27:22.629: INFO @log_variables: valid age_mae mean: 6.725943 2020-02-01 18:27:22.629: INFO @log_variables: valid gender_accuracy mean: 0.908328 2020-02-01 18:27:22.629: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055670 2020-02-01 18:27:22.629: INFO @log_variables: valid gender_confidence/accuracy mean: 0.851579 2020-02-01 18:27:22.629: INFO @log_variables: valid age_confidence/loss mean: 0.067973 2020-02-01 18:27:22.629: INFO @log_variables: valid age_confidence/accuracy mean: 0.560179 2020-02-01 18:27:22.629: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:27:22.636: INFO @metrics_hook: train age_mae: 6.948 +-0.040 (110372) 2020-02-01 18:27:22.643: INFO @metrics_hook: train gender_accuracy: 0.914 +-0.002 (110372) 2020-02-01 18:27:25.357: INFO @metrics_hook: valid age_mae: 6.726 +-0.094 (17639) 2020-02-01 18:27:25.358: INFO @metrics_hook: valid gender_accuracy: 0.908 +-0.004 (17639) 2020-02-01 18:27:26.991: INFO @decay_lr : LR updated to `8.2243234e-05` 2020-02-01 18:27:26.993: INFO @log_profile : T train: 128.164586 2020-02-01 18:27:26.993: INFO @log_profile : T valid: 5.709569 2020-02-01 18:27:26.993: INFO @log_profile : T read data: 2.830811 2020-02-01 18:27:26.993: INFO @log_profile : T hooks: 4.970884 2020-02-01 18:27:26.993: INFO @main_loop : Epoch 39 done 2020-02-01 18:27:26.993: INFO @main_loop : Training epoch 40 2020-02-01 18:29:40.815: INFO @log_variables: train loss nanmean: 0.894335 2020-02-01 18:29:40.815: INFO @log_variables: train age_loss mean: 6.423889 2020-02-01 18:29:40.815: INFO @log_variables: train gender_loss mean: 0.201150 2020-02-01 18:29:40.815: INFO @log_variables: train age_mae mean: 6.904347 2020-02-01 18:29:40.815: INFO @log_variables: train gender_accuracy mean: 0.916881 2020-02-01 18:29:40.815: INFO @log_variables: train gender_confidence/loss nanmean: 0.062576 2020-02-01 18:29:40.815: INFO @log_variables: train gender_confidence/accuracy mean: 0.810251 2020-02-01 18:29:40.815: INFO @log_variables: train age_confidence/loss mean: 0.064905 2020-02-01 18:29:40.816: INFO @log_variables: train age_confidence/accuracy mean: 0.611405 2020-02-01 18:29:40.816: INFO @log_variables: valid loss nanmean: 0.898942 2020-02-01 18:29:40.816: INFO @log_variables: valid age_loss mean: 6.124200 2020-02-01 18:29:40.816: INFO @log_variables: valid gender_loss mean: 0.238277 2020-02-01 18:29:40.816: INFO @log_variables: valid age_mae mean: 6.604537 2020-02-01 18:29:40.816: INFO @log_variables: valid gender_accuracy mean: 0.897727 2020-02-01 18:29:40.816: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057451 2020-02-01 18:29:40.816: INFO @log_variables: valid gender_confidence/accuracy mean: 0.859516 2020-02-01 18:29:40.816: INFO @log_variables: valid age_confidence/loss mean: 0.068130 2020-02-01 18:29:40.816: INFO @log_variables: valid age_confidence/accuracy mean: 0.562277 2020-02-01 18:29:40.816: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:29:40.823: INFO @metrics_hook: train age_mae: 6.904 +-0.039 (110372) 2020-02-01 18:29:40.831: INFO @metrics_hook: train gender_accuracy: 0.917 +-0.002 (110372) 2020-02-01 18:29:43.563: INFO @metrics_hook: valid age_mae: 6.605 +-0.093 (17639) 2020-02-01 18:29:43.564: INFO @metrics_hook: valid gender_accuracy: 0.898 +-0.005 (17639) 2020-02-01 18:29:45.021: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:29:45.021: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:29:45.022: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.19 2020-02-01 18:29:45.022: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.26 2020-02-01 18:29:45.154: INFO @evaluate_confidence: Previous accuracy would be: 91.69 2020-02-01 18:29:45.154: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:29:45.209: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.02, 96.15, 96.3, 96.41, 96.55, 96.69, 96.81, 96.92, 97.03, 97.11, 97.21, 97.31, 97.42, 97.51, 97.6, 97.68, 97.76, 97.84, 97.94, 98.03, 98.1, 98.19, 98.26, 98.33, 98.41, 98.47, 98.54, 98.62, 98.67, 98.73, 98.78, 98.83, 98.89, 98.94, 99.0] 2020-02-01 18:29:45.209: INFO @evaluate_confidence: Dropped ratios are: [17.5, 18.21, 18.96, 19.66, 20.43, 21.19, 21.87, 22.52, 23.21, 23.88, 24.57, 25.26, 25.95, 26.65, 27.3, 28.0, 28.66, 29.35, 30.02, 30.7, 31.4, 32.11, 32.79, 33.5, 34.2, 34.9, 35.65, 36.39, 37.13, 37.84, 38.66, 39.42, 40.25, 41.07, 41.92] 2020-02-01 18:29:45.259: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:29:45.259: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:29:45.259: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:29:45.259: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.14 2020-02-01 18:29:45.396: INFO @evaluate_confidence: Previous accuracy would be: 49.76 2020-02-01 18:29:45.396: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:29:45.414: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.46, 60.11, 60.8, 61.6, 62.43, 63.22, 64.05, 64.89, 65.57] 2020-02-01 18:29:45.415: INFO @evaluate_confidence: Dropped ratios are: [44.82, 47.67, 50.54, 53.34, 56.14, 58.93, 61.67, 64.22, 66.58] 2020-02-01 18:29:45.422: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:29:45.422: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.20 2020-02-01 18:29:45.422: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.20 2020-02-01 18:29:45.422: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.23 2020-02-01 18:29:45.520: INFO @evaluate_confidence: Previous accuracy would be: 89.77 2020-02-01 18:29:45.520: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 18:29:45.527: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.84, 95.02, 95.23, 95.4, 95.58, 95.75, 95.84, 95.92, 96.07, 96.21, 96.32, 96.41, 96.55, 96.66, 96.74, 96.84, 96.93, 97.11, 97.16, 97.24, 97.33, 97.41, 97.48, 97.56, 97.7, 97.86, 97.93, 98.0, 98.13, 98.17, 98.28, 98.39, 98.49] 2020-02-01 18:29:45.528: INFO @evaluate_confidence: Dropped ratios are: [15.57, 16.11, 16.79, 17.49, 18.08, 18.75, 19.28, 19.89, 20.51, 21.06, 21.67, 22.35, 22.9, 23.57, 24.19, 24.75, 25.34, 26.1, 26.8, 27.56, 28.31, 29.0, 29.55, 30.32, 31.13, 31.9, 32.76, 33.58, 34.49, 35.29, 36.2, 37.19, 38.23] 2020-02-01 18:29:45.535: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:29:45.535: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 18:29:45.535: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 18:29:45.536: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 18:29:45.662: INFO @evaluate_confidence: Previous accuracy would be: 50.82 2020-02-01 18:29:45.662: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 18:29:45.664: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.11, 59.46, 59.79, 60.3, 60.68, 60.65] 2020-02-01 18:29:45.664: INFO @evaluate_confidence: Dropped ratios are: [45.54, 49.93, 54.55, 58.89, 63.09, 66.95] 2020-02-01 18:29:45.714: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:29:46.395: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:29:46.479: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:29:46.930: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:29:47.003: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:29:47.682: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:29:47.768: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:29:47.770: INFO @evaluate_gender-age_model: groups 0 5.581535 1 5.715642 2 6.290544 3 6.233744 4 7.256975 5 7.512330 6 7.820055 7 9.438524 Name: errors, dtype: float64 2020-02-01 18:29:47.771: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:29:48.215: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:29:48.278: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:29:48.279: INFO @evaluate_gender-age_model: groups 0 7.454540 1 6.163589 2 6.448228 3 5.924400 4 7.385961 5 5.536488 6 7.876409 7 13.085775 Name: errors, dtype: float64 2020-02-01 18:29:48.433: INFO @decay_lr : LR updated to `8.183202e-05` 2020-02-01 18:29:48.435: INFO @log_profile : T train: 122.790919 2020-02-01 18:29:48.435: INFO @log_profile : T valid: 5.617732 2020-02-01 18:29:48.435: INFO @log_profile : T read data: 2.872060 2020-02-01 18:29:48.435: INFO @log_profile : T hooks: 10.085064 2020-02-01 18:29:48.435: INFO @main_loop : Epoch 40 done 2020-02-01 18:29:48.435: INFO @main_loop : Training epoch 41 2020-02-01 18:32:05.083: INFO @log_variables: train loss nanmean: 0.889368 2020-02-01 18:32:05.083: INFO @log_variables: train age_loss mean: 6.388234 2020-02-01 18:32:05.083: INFO @log_variables: train gender_loss mean: 0.199475 2020-02-01 18:32:05.083: INFO @log_variables: train age_mae mean: 6.868446 2020-02-01 18:32:05.083: INFO @log_variables: train gender_accuracy mean: 0.916278 2020-02-01 18:32:05.083: INFO @log_variables: train gender_confidence/loss nanmean: 0.062183 2020-02-01 18:32:05.083: INFO @log_variables: train gender_confidence/accuracy mean: 0.810655 2020-02-01 18:32:05.083: INFO @log_variables: train age_confidence/loss mean: 0.065096 2020-02-01 18:32:05.083: INFO @log_variables: train age_confidence/accuracy mean: 0.609619 2020-02-01 18:32:05.083: INFO @log_variables: valid loss nanmean: 0.889928 2020-02-01 18:32:05.084: INFO @log_variables: valid age_loss mean: 6.253655 2020-02-01 18:32:05.084: INFO @log_variables: valid gender_loss mean: 0.217625 2020-02-01 18:32:05.084: INFO @log_variables: valid age_mae mean: 6.733242 2020-02-01 18:32:05.084: INFO @log_variables: valid gender_accuracy mean: 0.904983 2020-02-01 18:32:05.084: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055127 2020-02-01 18:32:05.084: INFO @log_variables: valid gender_confidence/accuracy mean: 0.857418 2020-02-01 18:32:05.084: INFO @log_variables: valid age_confidence/loss mean: 0.068447 2020-02-01 18:32:05.084: INFO @log_variables: valid age_confidence/accuracy mean: 0.559896 2020-02-01 18:32:05.084: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:32:05.091: INFO @metrics_hook: train age_mae: 6.868 +-0.039 (110592) 2020-02-01 18:32:05.098: INFO @metrics_hook: train gender_accuracy: 0.916 +-0.002 (110592) 2020-02-01 18:32:07.796: INFO @metrics_hook: valid age_mae: 6.733 +-0.096 (17639) 2020-02-01 18:32:07.797: INFO @metrics_hook: valid gender_accuracy: 0.905 +-0.004 (17639) 2020-02-01 18:32:09.411: INFO @decay_lr : LR updated to `8.1422855e-05` 2020-02-01 18:32:09.412: INFO @log_profile : T train: 128.472270 2020-02-01 18:32:09.412: INFO @log_profile : T valid: 5.645679 2020-02-01 18:32:09.412: INFO @log_profile : T read data: 1.831137 2020-02-01 18:32:09.412: INFO @log_profile : T hooks: 4.952055 2020-02-01 18:32:09.412: INFO @main_loop : Epoch 41 done 2020-02-01 18:32:09.412: INFO @main_loop : Training epoch 42 2020-02-01 18:34:21.364: INFO @log_variables: train loss nanmean: 0.889841 2020-02-01 18:34:21.364: INFO @log_variables: train age_loss mean: 6.401510 2020-02-01 18:34:21.364: INFO @log_variables: train gender_loss mean: 0.198765 2020-02-01 18:34:21.365: INFO @log_variables: train age_mae mean: 6.882175 2020-02-01 18:34:21.365: INFO @log_variables: train gender_accuracy mean: 0.917588 2020-02-01 18:34:21.365: INFO @log_variables: train gender_confidence/loss nanmean: 0.062281 2020-02-01 18:34:21.365: INFO @log_variables: train gender_confidence/accuracy mean: 0.812561 2020-02-01 18:34:21.365: INFO @log_variables: train age_confidence/loss mean: 0.064910 2020-02-01 18:34:21.365: INFO @log_variables: train age_confidence/accuracy mean: 0.609684 2020-02-01 18:34:21.365: INFO @log_variables: valid loss nanmean: 0.875455 2020-02-01 18:34:21.365: INFO @log_variables: valid age_loss mean: 6.144098 2020-02-01 18:34:21.365: INFO @log_variables: valid gender_loss mean: 0.214448 2020-02-01 18:34:21.365: INFO @log_variables: valid age_mae mean: 6.623701 2020-02-01 18:34:21.365: INFO @log_variables: valid gender_accuracy mean: 0.908045 2020-02-01 18:34:21.365: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054518 2020-02-01 18:34:21.365: INFO @log_variables: valid gender_confidence/accuracy mean: 0.852032 2020-02-01 18:34:21.365: INFO @log_variables: valid age_confidence/loss mean: 0.067430 2020-02-01 18:34:21.365: INFO @log_variables: valid age_confidence/accuracy mean: 0.577754 2020-02-01 18:34:21.365: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:34:21.372: INFO @metrics_hook: train age_mae: 6.882 +-0.039 (110372) 2020-02-01 18:34:21.380: INFO @metrics_hook: train gender_accuracy: 0.918 +-0.002 (110372) 2020-02-01 18:34:24.070: INFO @metrics_hook: valid age_mae: 6.624 +-0.094 (17639) 2020-02-01 18:34:24.071: INFO @metrics_hook: valid gender_accuracy: 0.908 +-0.004 (17639) 2020-02-01 18:34:25.538: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:34:25.539: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:34:25.539: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.19 2020-02-01 18:34:25.539: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.25 2020-02-01 18:34:25.666: INFO @evaluate_confidence: Previous accuracy would be: 91.76 2020-02-01 18:34:25.666: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:34:25.719: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.08, 96.23, 96.35, 96.46, 96.58, 96.71, 96.81, 96.96, 97.05, 97.16, 97.26, 97.35, 97.44, 97.52, 97.61, 97.69, 97.78, 97.87, 97.96, 98.03, 98.1, 98.16, 98.23, 98.3, 98.37, 98.44, 98.5, 98.57, 98.63, 98.68, 98.75, 98.79, 98.86, 98.91, 98.97] 2020-02-01 18:34:25.719: INFO @evaluate_confidence: Dropped ratios are: [17.34, 18.06, 18.75, 19.45, 20.13, 20.86, 21.52, 22.25, 22.95, 23.63, 24.32, 25.01, 25.68, 26.39, 27.06, 27.71, 28.42, 29.1, 29.78, 30.47, 31.17, 31.88, 32.6, 33.33, 34.09, 34.84, 35.62, 36.34, 37.12, 37.87, 38.61, 39.4, 40.24, 41.04, 41.86] 2020-02-01 18:34:25.767: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:34:25.767: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:34:25.767: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:34:25.768: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.14 2020-02-01 18:34:25.901: INFO @evaluate_confidence: Previous accuracy would be: 49.73 2020-02-01 18:34:25.901: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:34:25.919: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.44, 60.11, 60.85, 61.5, 62.2, 62.89, 63.55, 64.42, 65.27] 2020-02-01 18:34:25.919: INFO @evaluate_confidence: Dropped ratios are: [44.63, 47.53, 50.38, 53.29, 56.13, 58.9, 61.43, 63.98, 66.45] 2020-02-01 18:34:25.926: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:34:25.927: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.22 2020-02-01 18:34:25.927: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.19 2020-02-01 18:34:25.927: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 18:34:26.022: INFO @evaluate_confidence: Previous accuracy would be: 90.80 2020-02-01 18:34:26.023: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 18:34:26.031: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.48, 95.75, 95.94, 96.06, 96.24, 96.4, 96.55, 96.71, 96.82, 96.89, 97.01, 97.13, 97.21, 97.3, 97.39, 97.54, 97.61, 97.71, 97.75, 97.81, 97.9, 97.97, 98.03, 98.08, 98.17, 98.26, 98.34, 98.38, 98.46, 98.53, 98.58, 98.65, 98.73, 98.77, 98.82, 98.87, 98.95, 99.04] 2020-02-01 18:34:26.031: INFO @evaluate_confidence: Dropped ratios are: [14.54, 15.23, 15.92, 16.53, 17.12, 17.73, 18.36, 18.96, 19.46, 20.06, 20.77, 21.47, 22.1, 22.78, 23.4, 24.08, 24.71, 25.36, 25.95, 26.64, 27.35, 28.14, 28.81, 29.47, 30.22, 31.07, 31.89, 32.64, 33.45, 34.28, 35.2, 35.98, 36.8, 37.64, 38.48, 39.34, 40.42, 41.34] 2020-02-01 18:34:26.038: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:34:26.038: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 18:34:26.039: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.10 2020-02-01 18:34:26.039: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 18:34:26.163: INFO @evaluate_confidence: Previous accuracy would be: 50.51 2020-02-01 18:34:26.163: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 18:34:26.165: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.86, 60.42, 60.86, 61.12, 61.32, 61.37, 61.61] 2020-02-01 18:34:26.165: INFO @evaluate_confidence: Dropped ratios are: [42.07, 45.42, 49.19, 52.81, 56.85, 60.85, 64.32] 2020-02-01 18:34:26.216: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:34:26.900: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:34:26.985: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:34:27.414: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:34:27.640: INFO @decay_lr : LR updated to `8.101574e-05` 2020-02-01 18:34:27.641: INFO @log_profile : T train: 122.917586 2020-02-01 18:34:27.641: INFO @log_profile : T valid: 5.558868 2020-02-01 18:34:27.641: INFO @log_profile : T read data: 2.787813 2020-02-01 18:34:27.641: INFO @log_profile : T hooks: 6.886926 2020-02-01 18:34:27.641: INFO @main_loop : Epoch 42 done 2020-02-01 18:34:27.642: INFO @main_loop : Training epoch 43 2020-02-01 18:36:47.443: INFO @log_variables: train loss nanmean: 0.879958 2020-02-01 18:36:47.444: INFO @log_variables: train age_loss mean: 6.332577 2020-02-01 18:36:47.444: INFO @log_variables: train gender_loss mean: 0.194954 2020-02-01 18:36:47.444: INFO @log_variables: train age_mae mean: 6.813121 2020-02-01 18:36:47.444: INFO @log_variables: train gender_accuracy mean: 0.918847 2020-02-01 18:36:47.444: INFO @log_variables: train gender_confidence/loss nanmean: 0.061844 2020-02-01 18:36:47.444: INFO @log_variables: train gender_confidence/accuracy mean: 0.812471 2020-02-01 18:36:47.444: INFO @log_variables: train age_confidence/loss mean: 0.065195 2020-02-01 18:36:47.444: INFO @log_variables: train age_confidence/accuracy mean: 0.607681 2020-02-01 18:36:47.444: INFO @log_variables: valid loss nanmean: 0.883218 2020-02-01 18:36:47.444: INFO @log_variables: valid age_loss mean: 6.066967 2020-02-01 18:36:47.444: INFO @log_variables: valid gender_loss mean: 0.225892 2020-02-01 18:36:47.444: INFO @log_variables: valid age_mae mean: 6.547435 2020-02-01 18:36:47.444: INFO @log_variables: valid gender_accuracy mean: 0.903849 2020-02-01 18:36:47.444: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057529 2020-02-01 18:36:47.444: INFO @log_variables: valid gender_confidence/accuracy mean: 0.858382 2020-02-01 18:36:47.444: INFO @log_variables: valid age_confidence/loss mean: 0.068790 2020-02-01 18:36:47.444: INFO @log_variables: valid age_confidence/accuracy mean: 0.574692 2020-02-01 18:36:47.445: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:36:47.452: INFO @metrics_hook: train age_mae: 6.813 +-0.039 (110372) 2020-02-01 18:36:47.459: INFO @metrics_hook: train gender_accuracy: 0.919 +-0.002 (110372) 2020-02-01 18:36:50.238: INFO @metrics_hook: valid age_mae: 6.547 +-0.093 (17639) 2020-02-01 18:36:50.239: INFO @metrics_hook: valid gender_accuracy: 0.904 +-0.004 (17639) 2020-02-01 18:36:51.891: INFO @decay_lr : LR updated to `8.061066e-05` 2020-02-01 18:36:51.893: INFO @log_profile : T train: 130.049982 2020-02-01 18:36:51.893: INFO @log_profile : T valid: 6.364657 2020-02-01 18:36:51.893: INFO @log_profile : T read data: 2.715084 2020-02-01 18:36:51.893: INFO @log_profile : T hooks: 5.047287 2020-02-01 18:36:51.893: INFO @main_loop : Epoch 43 done 2020-02-01 18:36:51.893: INFO @main_loop : Training epoch 44 2020-02-01 18:39:11.943: INFO @log_variables: train loss nanmean: 0.871682 2020-02-01 18:39:11.943: INFO @log_variables: train age_loss mean: 6.256427 2020-02-01 18:39:11.943: INFO @log_variables: train gender_loss mean: 0.193156 2020-02-01 18:39:11.943: INFO @log_variables: train age_mae mean: 6.736273 2020-02-01 18:39:11.943: INFO @log_variables: train gender_accuracy mean: 0.919904 2020-02-01 18:39:11.943: INFO @log_variables: train gender_confidence/loss nanmean: 0.061894 2020-02-01 18:39:11.943: INFO @log_variables: train gender_confidence/accuracy mean: 0.816614 2020-02-01 18:39:11.943: INFO @log_variables: train age_confidence/loss mean: 0.065425 2020-02-01 18:39:11.943: INFO @log_variables: train age_confidence/accuracy mean: 0.608724 2020-02-01 18:39:11.943: INFO @log_variables: valid loss nanmean: 0.860587 2020-02-01 18:39:11.943: INFO @log_variables: valid age_loss mean: 6.015031 2020-02-01 18:39:11.943: INFO @log_variables: valid gender_loss mean: 0.207262 2020-02-01 18:39:11.943: INFO @log_variables: valid age_mae mean: 6.494871 2020-02-01 18:39:11.943: INFO @log_variables: valid gender_accuracy mean: 0.911163 2020-02-01 18:39:11.943: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055506 2020-02-01 18:39:11.944: INFO @log_variables: valid gender_confidence/accuracy mean: 0.854754 2020-02-01 18:39:11.944: INFO @log_variables: valid age_confidence/loss mean: 0.069841 2020-02-01 18:39:11.944: INFO @log_variables: valid age_confidence/accuracy mean: 0.547593 2020-02-01 18:39:11.944: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:39:11.951: INFO @metrics_hook: train age_mae: 6.736 +-0.038 (110592) 2020-02-01 18:39:11.958: INFO @metrics_hook: train gender_accuracy: 0.920 +-0.002 (110592) 2020-02-01 18:39:14.676: INFO @metrics_hook: valid age_mae: 6.495 +-0.094 (17639) 2020-02-01 18:39:14.677: INFO @metrics_hook: valid gender_accuracy: 0.911 +-0.004 (17639) 2020-02-01 18:39:16.146: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:39:16.147: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:39:16.147: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.19 2020-02-01 18:39:16.147: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.26 2020-02-01 18:39:16.281: INFO @evaluate_confidence: Previous accuracy would be: 91.99 2020-02-01 18:39:16.281: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:39:16.337: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.15, 96.3, 96.42, 96.54, 96.66, 96.78, 96.9, 96.98, 97.09, 97.2, 97.3, 97.39, 97.46, 97.55, 97.65, 97.73, 97.82, 97.91, 97.98, 98.05, 98.12, 98.21, 98.28, 98.35, 98.41, 98.48, 98.52, 98.58, 98.63, 98.69, 98.74, 98.8, 98.85, 98.91, 98.95, 98.99] 2020-02-01 18:39:16.337: INFO @evaluate_confidence: Dropped ratios are: [16.44, 17.11, 17.75, 18.44, 19.12, 19.76, 20.5, 21.15, 21.8, 22.45, 23.14, 23.8, 24.42, 25.1, 25.77, 26.41, 27.13, 27.84, 28.53, 29.18, 29.87, 30.55, 31.25, 31.99, 32.67, 33.36, 34.03, 34.78, 35.52, 36.24, 36.99, 37.77, 38.55, 39.41, 40.24, 41.06] 2020-02-01 18:39:16.386: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:39:16.387: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:39:16.387: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:39:16.387: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.14 2020-02-01 18:39:16.527: INFO @evaluate_confidence: Previous accuracy would be: 50.63 2020-02-01 18:39:16.527: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:39:16.546: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [60.12, 60.82, 61.4, 62.13, 62.76, 63.47, 64.24, 65.07, 65.86] 2020-02-01 18:39:16.546: INFO @evaluate_confidence: Dropped ratios are: [43.44, 46.32, 49.14, 52.09, 54.93, 57.68, 60.41, 63.08, 65.54] 2020-02-01 18:39:16.554: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:39:16.554: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.21 2020-02-01 18:39:16.554: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.20 2020-02-01 18:39:16.554: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.24 2020-02-01 18:39:16.650: INFO @evaluate_confidence: Previous accuracy would be: 91.12 2020-02-01 18:39:16.651: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 18:39:16.659: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.71, 95.82, 95.93, 96.1, 96.18, 96.31, 96.44, 96.53, 96.61, 96.75, 96.86, 96.94, 97.04, 97.18, 97.25, 97.39, 97.53, 97.58, 97.66, 97.76, 97.78, 97.84, 97.92, 97.96, 98.01, 98.06, 98.15, 98.24, 98.28, 98.39, 98.47, 98.53, 98.61, 98.67, 98.72, 98.75] 2020-02-01 18:39:16.659: INFO @evaluate_confidence: Dropped ratios are: [14.73, 15.27, 15.77, 16.54, 17.08, 17.62, 18.23, 18.8, 19.33, 19.81, 20.4, 21.0, 21.51, 22.13, 22.76, 23.45, 24.17, 24.71, 25.34, 25.95, 26.67, 27.25, 27.93, 28.65, 29.35, 30.06, 30.84, 31.41, 32.24, 33.03, 33.95, 34.82, 35.67, 36.6, 37.54, 38.55] 2020-02-01 18:39:16.667: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:39:16.667: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.11 2020-02-01 18:39:16.667: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 18:39:16.667: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 18:39:16.799: INFO @evaluate_confidence: Previous accuracy would be: 52.54 2020-02-01 18:39:16.799: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 18:39:16.801: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [60.24, 60.4, 60.52, 60.58, 60.93] 2020-02-01 18:39:16.801: INFO @evaluate_confidence: Dropped ratios are: [47.19, 51.27, 55.39, 59.32, 63.01] 2020-02-01 18:39:16.853: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:39:17.544: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:39:17.629: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:39:18.084: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:39:18.316: INFO @decay_lr : LR updated to `8.020761e-05` 2020-02-01 18:39:18.628: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 18:39:18.631: INFO @log_profile : T train: 130.654396 2020-02-01 18:39:18.631: INFO @log_profile : T valid: 6.864419 2020-02-01 18:39:18.631: INFO @log_profile : T read data: 1.850173 2020-02-01 18:39:18.631: INFO @log_profile : T hooks: 7.292925 2020-02-01 18:39:18.631: INFO @main_loop : Epoch 44 done 2020-02-01 18:39:18.631: INFO @main_loop : Training epoch 45 2020-02-01 18:41:36.966: INFO @log_variables: train loss nanmean: 0.867155 2020-02-01 18:41:36.966: INFO @log_variables: train age_loss mean: 6.227621 2020-02-01 18:41:36.966: INFO @log_variables: train gender_loss mean: 0.190954 2020-02-01 18:41:36.966: INFO @log_variables: train age_mae mean: 6.708246 2020-02-01 18:41:36.966: INFO @log_variables: train gender_accuracy mean: 0.920723 2020-02-01 18:41:36.966: INFO @log_variables: train gender_confidence/loss nanmean: 0.061603 2020-02-01 18:41:36.966: INFO @log_variables: train gender_confidence/accuracy mean: 0.816847 2020-02-01 18:41:36.966: INFO @log_variables: train age_confidence/loss mean: 0.065810 2020-02-01 18:41:36.966: INFO @log_variables: train age_confidence/accuracy mean: 0.606766 2020-02-01 18:41:36.966: INFO @log_variables: valid loss nanmean: 0.903800 2020-02-01 18:41:36.966: INFO @log_variables: valid age_loss mean: 6.122314 2020-02-01 18:41:36.966: INFO @log_variables: valid gender_loss mean: 0.239293 2020-02-01 18:41:36.966: INFO @log_variables: valid age_mae mean: 6.603586 2020-02-01 18:41:36.967: INFO @log_variables: valid gender_accuracy mean: 0.903793 2020-02-01 18:41:36.967: INFO @log_variables: valid gender_confidence/loss nanmean: 0.060552 2020-02-01 18:41:36.967: INFO @log_variables: valid gender_confidence/accuracy mean: 0.848177 2020-02-01 18:41:36.967: INFO @log_variables: valid age_confidence/loss mean: 0.069134 2020-02-01 18:41:36.967: INFO @log_variables: valid age_confidence/accuracy mean: 0.541471 2020-02-01 18:41:36.967: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:41:36.974: INFO @metrics_hook: train age_mae: 6.708 +-0.038 (110372) 2020-02-01 18:41:36.982: INFO @metrics_hook: train gender_accuracy: 0.921 +-0.002 (110372) 2020-02-01 18:41:39.715: INFO @metrics_hook: valid age_mae: 6.604 +-0.093 (17639) 2020-02-01 18:41:39.717: INFO @metrics_hook: valid gender_accuracy: 0.904 +-0.004 (17639) 2020-02-01 18:41:43.994: INFO @decay_lr : LR updated to `7.9806574e-05` 2020-02-01 18:41:43.996: INFO @log_profile : T train: 129.252799 2020-02-01 18:41:43.996: INFO @log_profile : T valid: 5.627514 2020-02-01 18:41:43.996: INFO @log_profile : T read data: 2.778416 2020-02-01 18:41:43.996: INFO @log_profile : T hooks: 7.629615 2020-02-01 18:41:43.996: INFO @main_loop : Epoch 45 done 2020-02-01 18:41:43.996: INFO @main_loop : Training epoch 46 2020-02-01 18:43:55.702: INFO @log_variables: train loss nanmean: 0.869103 2020-02-01 18:43:55.703: INFO @log_variables: train age_loss mean: 6.238132 2020-02-01 18:43:55.703: INFO @log_variables: train gender_loss mean: 0.191967 2020-02-01 18:43:55.703: INFO @log_variables: train age_mae mean: 6.717875 2020-02-01 18:43:55.703: INFO @log_variables: train gender_accuracy mean: 0.920596 2020-02-01 18:43:55.703: INFO @log_variables: train gender_confidence/loss nanmean: 0.061808 2020-02-01 18:43:55.703: INFO @log_variables: train gender_confidence/accuracy mean: 0.815252 2020-02-01 18:43:55.703: INFO @log_variables: train age_confidence/loss mean: 0.065676 2020-02-01 18:43:55.703: INFO @log_variables: train age_confidence/accuracy mean: 0.606032 2020-02-01 18:43:55.703: INFO @log_variables: valid loss nanmean: 0.891356 2020-02-01 18:43:55.703: INFO @log_variables: valid age_loss mean: 6.051744 2020-02-01 18:43:55.703: INFO @log_variables: valid gender_loss mean: 0.233929 2020-02-01 18:43:55.703: INFO @log_variables: valid age_mae mean: 6.532056 2020-02-01 18:43:55.703: INFO @log_variables: valid gender_accuracy mean: 0.902886 2020-02-01 18:43:55.703: INFO @log_variables: valid gender_confidence/loss nanmean: 0.059899 2020-02-01 18:43:55.703: INFO @log_variables: valid gender_confidence/accuracy mean: 0.859743 2020-02-01 18:43:55.703: INFO @log_variables: valid age_confidence/loss mean: 0.068635 2020-02-01 18:43:55.703: INFO @log_variables: valid age_confidence/accuracy mean: 0.559329 2020-02-01 18:43:55.704: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:43:55.711: INFO @metrics_hook: train age_mae: 6.718 +-0.038 (110372) 2020-02-01 18:43:55.718: INFO @metrics_hook: train gender_accuracy: 0.921 +-0.002 (110372) 2020-02-01 18:43:58.456: INFO @metrics_hook: valid age_mae: 6.532 +-0.091 (17639) 2020-02-01 18:43:58.458: INFO @metrics_hook: valid gender_accuracy: 0.903 +-0.004 (17639) 2020-02-01 18:43:59.899: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:43:59.899: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:43:59.899: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.19 2020-02-01 18:43:59.900: INFO @evaluate_confidence: Average confidence of all samples 0.75 +- 0.26 2020-02-01 18:44:00.028: INFO @evaluate_confidence: Previous accuracy would be: 92.06 2020-02-01 18:44:00.028: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:44:00.083: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.18, 96.31, 96.45, 96.55, 96.66, 96.78, 96.87, 96.98, 97.09, 97.19, 97.3, 97.39, 97.5, 97.57, 97.66, 97.75, 97.84, 97.93, 98.0, 98.07, 98.15, 98.22, 98.3, 98.37, 98.44, 98.49, 98.54, 98.6, 98.66, 98.74, 98.78, 98.84, 98.88, 98.92, 98.96, 98.99] 2020-02-01 18:44:00.083: INFO @evaluate_confidence: Dropped ratios are: [16.46, 17.15, 17.87, 18.5, 19.14, 19.8, 20.45, 21.16, 21.86, 22.56, 23.22, 23.88, 24.58, 25.21, 25.88, 26.52, 27.13, 27.86, 28.52, 29.18, 29.81, 30.49, 31.17, 31.85, 32.54, 33.23, 33.92, 34.61, 35.35, 36.12, 36.86, 37.65, 38.42, 39.18, 40.02, 40.88] 2020-02-01 18:44:00.131: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:44:00.131: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:44:00.131: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:44:00.132: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.14 2020-02-01 18:44:00.266: INFO @evaluate_confidence: Previous accuracy would be: 50.77 2020-02-01 18:44:00.266: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:44:00.284: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.93, 60.57, 61.19, 61.92, 62.52, 63.31, 64.07, 64.89, 65.7] 2020-02-01 18:44:00.285: INFO @evaluate_confidence: Dropped ratios are: [42.88, 45.85, 48.77, 51.72, 54.59, 57.36, 60.13, 62.78, 65.25] 2020-02-01 18:44:00.292: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:44:00.292: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.21 2020-02-01 18:44:00.292: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.21 2020-02-01 18:44:00.292: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.23 2020-02-01 18:44:00.387: INFO @evaluate_confidence: Previous accuracy would be: 90.29 2020-02-01 18:44:00.387: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 18:44:00.394: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.98, 95.11, 95.3, 95.43, 95.53, 95.63, 95.76, 95.93, 96.04, 96.14, 96.23, 96.35, 96.5, 96.59, 96.67, 96.76, 96.86, 96.93, 97.01, 97.13, 97.19, 97.32, 97.41, 97.49, 97.55, 97.63, 97.67, 97.75, 97.81, 97.92, 97.99, 98.06, 98.16] 2020-02-01 18:44:00.394: INFO @evaluate_confidence: Dropped ratios are: [15.22, 15.75, 16.37, 16.99, 17.5, 18.02, 18.64, 19.35, 19.96, 20.55, 21.2, 21.89, 22.56, 23.28, 23.94, 24.5, 25.22, 25.84, 26.5, 27.34, 28.13, 28.8, 29.42, 30.19, 31.05, 31.88, 32.68, 33.53, 34.42, 35.38, 36.33, 37.3, 38.15] 2020-02-01 18:44:00.402: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:44:00.402: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 18:44:00.402: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 18:44:00.402: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 18:44:00.528: INFO @evaluate_confidence: Previous accuracy would be: 51.20 2020-02-01 18:44:00.528: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 18:44:00.529: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.61, 58.97, 59.09, 59.28] 2020-02-01 18:44:00.530: INFO @evaluate_confidence: Dropped ratios are: [44.92, 48.8, 52.47, 56.71] 2020-02-01 18:44:00.582: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:44:01.264: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:44:01.346: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:44:01.801: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:44:02.024: INFO @decay_lr : LR updated to `7.940754e-05` 2020-02-01 18:44:02.025: INFO @log_profile : T train: 122.848856 2020-02-01 18:44:02.025: INFO @log_profile : T valid: 5.408704 2020-02-01 18:44:02.025: INFO @log_profile : T read data: 2.760830 2020-02-01 18:44:02.025: INFO @log_profile : T hooks: 6.932660 2020-02-01 18:44:02.025: INFO @main_loop : Epoch 46 done 2020-02-01 18:44:02.025: INFO @main_loop : Training epoch 47 2020-02-01 18:46:21.757: INFO @log_variables: train loss nanmean: 0.860041 2020-02-01 18:46:21.757: INFO @log_variables: train age_loss mean: 6.177234 2020-02-01 18:46:21.757: INFO @log_variables: train gender_loss mean: 0.188612 2020-02-01 18:46:21.757: INFO @log_variables: train age_mae mean: 6.656681 2020-02-01 18:46:21.757: INFO @log_variables: train gender_accuracy mean: 0.921206 2020-02-01 18:46:21.757: INFO @log_variables: train gender_confidence/loss nanmean: 0.061160 2020-02-01 18:46:21.757: INFO @log_variables: train gender_confidence/accuracy mean: 0.818106 2020-02-01 18:46:21.757: INFO @log_variables: train age_confidence/loss mean: 0.065849 2020-02-01 18:46:21.758: INFO @log_variables: train age_confidence/accuracy mean: 0.609999 2020-02-01 18:46:21.758: INFO @log_variables: valid loss nanmean: 0.873934 2020-02-01 18:46:21.758: INFO @log_variables: valid age_loss mean: 6.019936 2020-02-01 18:46:21.758: INFO @log_variables: valid gender_loss mean: 0.218764 2020-02-01 18:46:21.758: INFO @log_variables: valid age_mae mean: 6.499312 2020-02-01 18:46:21.758: INFO @log_variables: valid gender_accuracy mean: 0.908328 2020-02-01 18:46:21.758: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058971 2020-02-01 18:46:21.758: INFO @log_variables: valid gender_confidence/accuracy mean: 0.859176 2020-02-01 18:46:21.758: INFO @log_variables: valid age_confidence/loss mean: 0.068820 2020-02-01 18:46:21.758: INFO @log_variables: valid age_confidence/accuracy mean: 0.562333 2020-02-01 18:46:21.758: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:46:21.765: INFO @metrics_hook: train age_mae: 6.657 +-0.038 (110592) 2020-02-01 18:46:21.772: INFO @metrics_hook: train gender_accuracy: 0.921 +-0.002 (110592) 2020-02-01 18:46:24.489: INFO @metrics_hook: valid age_mae: 6.499 +-0.092 (17639) 2020-02-01 18:46:24.490: INFO @metrics_hook: valid gender_accuracy: 0.908 +-0.004 (17639) 2020-02-01 18:46:26.122: INFO @decay_lr : LR updated to `7.9010504e-05` 2020-02-01 18:46:26.124: INFO @log_profile : T train: 130.425159 2020-02-01 18:46:26.124: INFO @log_profile : T valid: 6.753974 2020-02-01 18:46:26.124: INFO @log_profile : T read data: 1.873450 2020-02-01 18:46:26.124: INFO @log_profile : T hooks: 4.970438 2020-02-01 18:46:26.124: INFO @main_loop : Epoch 47 done 2020-02-01 18:46:26.124: INFO @main_loop : Training epoch 48 2020-02-01 18:48:45.475: INFO @log_variables: train loss nanmean: 0.853284 2020-02-01 18:48:45.475: INFO @log_variables: train age_loss mean: 6.154498 2020-02-01 18:48:45.475: INFO @log_variables: train gender_loss mean: 0.183434 2020-02-01 18:48:45.475: INFO @log_variables: train age_mae mean: 6.634042 2020-02-01 18:48:45.475: INFO @log_variables: train gender_accuracy mean: 0.924555 2020-02-01 18:48:45.475: INFO @log_variables: train gender_confidence/loss nanmean: 0.061090 2020-02-01 18:48:45.476: INFO @log_variables: train gender_confidence/accuracy mean: 0.820027 2020-02-01 18:48:45.476: INFO @log_variables: train age_confidence/loss mean: 0.065936 2020-02-01 18:48:45.476: INFO @log_variables: train age_confidence/accuracy mean: 0.607337 2020-02-01 18:48:45.476: INFO @log_variables: valid loss nanmean: 0.882581 2020-02-01 18:48:45.476: INFO @log_variables: valid age_loss mean: 6.216990 2020-02-01 18:48:45.476: INFO @log_variables: valid gender_loss mean: 0.214316 2020-02-01 18:48:45.476: INFO @log_variables: valid age_mae mean: 6.699030 2020-02-01 18:48:45.476: INFO @log_variables: valid gender_accuracy mean: 0.909292 2020-02-01 18:48:45.476: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054894 2020-02-01 18:48:45.476: INFO @log_variables: valid gender_confidence/accuracy mean: 0.854017 2020-02-01 18:48:45.476: INFO @log_variables: valid age_confidence/loss mean: 0.067672 2020-02-01 18:48:45.476: INFO @log_variables: valid age_confidence/accuracy mean: 0.557118 2020-02-01 18:48:45.476: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:48:45.483: INFO @metrics_hook: train age_mae: 6.634 +-0.038 (110372) 2020-02-01 18:48:45.490: INFO @metrics_hook: train gender_accuracy: 0.925 +-0.002 (110372) 2020-02-01 18:48:48.209: INFO @metrics_hook: valid age_mae: 6.699 +-0.093 (17639) 2020-02-01 18:48:48.210: INFO @metrics_hook: valid gender_accuracy: 0.909 +-0.004 (17639) 2020-02-01 18:48:49.631: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:48:49.632: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.78 +- 0.24 2020-02-01 18:48:49.632: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.20 2020-02-01 18:48:49.632: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 18:48:49.768: INFO @evaluate_confidence: Previous accuracy would be: 92.46 2020-02-01 18:48:49.768: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77] 2020-02-01 18:48:49.826: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.45, 96.56, 96.68, 96.78, 96.9, 97.01, 97.11, 97.21, 97.32, 97.42, 97.51, 97.62, 97.7, 97.79, 97.87, 97.94, 98.02, 98.07, 98.15, 98.22, 98.3, 98.35, 98.41, 98.46, 98.52, 98.57, 98.63, 98.68, 98.73, 98.77, 98.81, 98.87, 98.91, 98.95, 99.0, 99.04] 2020-02-01 18:48:49.826: INFO @evaluate_confidence: Dropped ratios are: [16.13, 16.75, 17.4, 18.05, 18.7, 19.4, 20.05, 20.66, 21.32, 21.96, 22.59, 23.24, 23.9, 24.58, 25.22, 25.86, 26.51, 27.13, 27.75, 28.46, 29.14, 29.77, 30.48, 31.12, 31.81, 32.48, 33.19, 33.91, 34.66, 35.4, 36.17, 36.96, 37.78, 38.59, 39.4, 40.26] 2020-02-01 18:48:49.877: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:48:49.877: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:48:49.877: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:48:49.877: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.14 2020-02-01 18:48:50.019: INFO @evaluate_confidence: Previous accuracy would be: 51.08 2020-02-01 18:48:50.020: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:48:50.038: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [60.31, 60.94, 61.6, 62.29, 62.8, 63.46, 64.2, 64.94, 65.68] 2020-02-01 18:48:50.038: INFO @evaluate_confidence: Dropped ratios are: [42.45, 45.32, 48.21, 51.1, 53.85, 56.62, 59.29, 61.88, 64.39] 2020-02-01 18:48:50.046: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:48:50.046: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.22 2020-02-01 18:48:50.046: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.20 2020-02-01 18:48:50.046: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 18:48:50.147: INFO @evaluate_confidence: Previous accuracy would be: 90.93 2020-02-01 18:48:50.147: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 18:48:50.155: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.42, 95.6, 95.78, 95.96, 96.12, 96.31, 96.45, 96.57, 96.67, 96.83, 96.94, 97.06, 97.17, 97.25, 97.37, 97.47, 97.62, 97.74, 97.84, 97.9, 97.99, 98.04, 98.07, 98.14, 98.17, 98.21, 98.25, 98.29, 98.33, 98.41, 98.52, 98.58, 98.67, 98.7, 98.74, 98.79, 98.89] 2020-02-01 18:48:50.155: INFO @evaluate_confidence: Dropped ratios are: [14.91, 15.43, 15.95, 16.46, 16.99, 17.58, 18.02, 18.6, 19.08, 19.66, 20.19, 20.81, 21.42, 21.96, 22.52, 23.07, 23.68, 24.37, 25.03, 25.61, 26.23, 26.93, 27.59, 28.31, 28.95, 29.66, 30.31, 31.09, 31.91, 32.8, 33.69, 34.46, 35.32, 36.11, 37.19, 38.3, 39.2] 2020-02-01 18:48:50.163: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:48:50.163: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 18:48:50.163: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.09 2020-02-01 18:48:50.163: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.10 2020-02-01 18:48:50.287: INFO @evaluate_confidence: Previous accuracy would be: 49.77 2020-02-01 18:48:50.287: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 18:48:50.289: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.47, 57.07, 57.31, 57.73, 58.0, 57.95] 2020-02-01 18:48:50.289: INFO @evaluate_confidence: Dropped ratios are: [44.17, 48.6, 53.09, 57.37, 61.5, 65.58] 2020-02-01 18:48:50.345: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:48:51.017: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:48:51.099: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:48:51.538: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:48:51.764: INFO @decay_lr : LR updated to `7.861545e-05` 2020-02-01 18:48:51.765: INFO @log_profile : T train: 130.183938 2020-02-01 18:48:51.765: INFO @log_profile : T valid: 5.708251 2020-02-01 18:48:51.765: INFO @log_profile : T read data: 2.807249 2020-02-01 18:48:51.765: INFO @log_profile : T hooks: 6.866337 2020-02-01 18:48:51.765: INFO @main_loop : Epoch 48 done 2020-02-01 18:48:51.766: INFO @main_loop : Training epoch 49 2020-02-01 18:51:08.498: INFO @log_variables: train loss nanmean: 0.854248 2020-02-01 18:51:08.498: INFO @log_variables: train age_loss mean: 6.131145 2020-02-01 18:51:08.498: INFO @log_variables: train gender_loss mean: 0.186286 2020-02-01 18:51:08.498: INFO @log_variables: train age_mae mean: 6.610525 2020-02-01 18:51:08.498: INFO @log_variables: train gender_accuracy mean: 0.923812 2020-02-01 18:51:08.498: INFO @log_variables: train gender_confidence/loss nanmean: 0.061458 2020-02-01 18:51:08.498: INFO @log_variables: train gender_confidence/accuracy mean: 0.816955 2020-02-01 18:51:08.498: INFO @log_variables: train age_confidence/loss mean: 0.066062 2020-02-01 18:51:08.498: INFO @log_variables: train age_confidence/accuracy mean: 0.608098 2020-02-01 18:51:08.498: INFO @log_variables: valid loss nanmean: 0.885799 2020-02-01 18:51:08.499: INFO @log_variables: valid age_loss mean: 6.017047 2020-02-01 18:51:08.499: INFO @log_variables: valid gender_loss mean: 0.234783 2020-02-01 18:51:08.499: INFO @log_variables: valid age_mae mean: 6.498296 2020-02-01 18:51:08.499: INFO @log_variables: valid gender_accuracy mean: 0.900051 2020-02-01 18:51:08.499: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056974 2020-02-01 18:51:08.499: INFO @log_variables: valid gender_confidence/accuracy mean: 0.849708 2020-02-01 18:51:08.499: INFO @log_variables: valid age_confidence/loss mean: 0.068382 2020-02-01 18:51:08.499: INFO @log_variables: valid age_confidence/accuracy mean: 0.555984 2020-02-01 18:51:08.499: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:51:08.506: INFO @metrics_hook: train age_mae: 6.611 +-0.038 (110372) 2020-02-01 18:51:08.514: INFO @metrics_hook: train gender_accuracy: 0.924 +-0.002 (110372) 2020-02-01 18:51:11.227: INFO @metrics_hook: valid age_mae: 6.498 +-0.090 (17639) 2020-02-01 18:51:11.228: INFO @metrics_hook: valid gender_accuracy: 0.900 +-0.005 (17639) 2020-02-01 18:51:12.859: INFO @decay_lr : LR updated to `7.822237e-05` 2020-02-01 18:51:12.861: INFO @log_profile : T train: 127.709982 2020-02-01 18:51:12.861: INFO @log_profile : T valid: 5.505610 2020-02-01 18:51:12.861: INFO @log_profile : T read data: 2.822784 2020-02-01 18:51:12.861: INFO @log_profile : T hooks: 4.982019 2020-02-01 18:51:12.861: INFO @main_loop : Epoch 49 done 2020-02-01 18:51:12.861: INFO @main_loop : Training epoch 50 2020-02-01 18:53:24.605: INFO @log_variables: train loss nanmean: 0.848350 2020-02-01 18:53:24.606: INFO @log_variables: train age_loss mean: 6.126221 2020-02-01 18:53:24.606: INFO @log_variables: train gender_loss mean: 0.181131 2020-02-01 18:53:24.606: INFO @log_variables: train age_mae mean: 6.605659 2020-02-01 18:53:24.606: INFO @log_variables: train gender_accuracy mean: 0.925067 2020-02-01 18:53:24.606: INFO @log_variables: train gender_confidence/loss nanmean: 0.060479 2020-02-01 18:53:24.606: INFO @log_variables: train gender_confidence/accuracy mean: 0.821967 2020-02-01 18:53:24.606: INFO @log_variables: train age_confidence/loss mean: 0.066277 2020-02-01 18:53:24.606: INFO @log_variables: train age_confidence/accuracy mean: 0.606988 2020-02-01 18:53:24.606: INFO @log_variables: valid loss nanmean: 0.901918 2020-02-01 18:53:24.606: INFO @log_variables: valid age_loss mean: 5.992579 2020-02-01 18:53:24.606: INFO @log_variables: valid gender_loss mean: 0.253665 2020-02-01 18:53:24.606: INFO @log_variables: valid age_mae mean: 6.472572 2020-02-01 18:53:24.606: INFO @log_variables: valid gender_accuracy mean: 0.893305 2020-02-01 18:53:24.606: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057095 2020-02-01 18:53:24.606: INFO @log_variables: valid gender_confidence/accuracy mean: 0.844209 2020-02-01 18:53:24.606: INFO @log_variables: valid age_confidence/loss mean: 0.069439 2020-02-01 18:53:24.606: INFO @log_variables: valid age_confidence/accuracy mean: 0.543625 2020-02-01 18:53:24.606: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:53:24.614: INFO @metrics_hook: train age_mae: 6.606 +-0.038 (110592) 2020-02-01 18:53:24.621: INFO @metrics_hook: train gender_accuracy: 0.925 +-0.002 (110592) 2020-02-01 18:53:27.365: INFO @metrics_hook: valid age_mae: 6.473 +-0.092 (17639) 2020-02-01 18:53:27.366: INFO @metrics_hook: valid gender_accuracy: 0.893 +-0.005 (17639) 2020-02-01 18:53:28.816: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:53:28.816: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 18:53:28.816: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.19 2020-02-01 18:53:28.817: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 18:53:28.946: INFO @evaluate_confidence: Previous accuracy would be: 92.51 2020-02-01 18:53:28.947: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 18:53:29.004: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.39, 96.52, 96.62, 96.74, 96.86, 96.98, 97.07, 97.17, 97.28, 97.37, 97.47, 97.57, 97.65, 97.74, 97.84, 97.93, 98.01, 98.09, 98.16, 98.22, 98.28, 98.34, 98.41, 98.46, 98.53, 98.59, 98.65, 98.71, 98.75, 98.82, 98.87, 98.91, 98.95, 99.01, 99.05, 99.1, 99.13, 99.17, 99.24] 2020-02-01 18:53:29.005: INFO @evaluate_confidence: Dropped ratios are: [15.45, 16.09, 16.72, 17.32, 17.94, 18.59, 19.23, 19.86, 20.52, 21.14, 21.76, 22.37, 23.0, 23.66, 24.29, 24.91, 25.55, 26.17, 26.8, 27.4, 28.02, 28.69, 29.35, 30.02, 30.74, 31.39, 32.04, 32.71, 33.41, 34.16, 34.88, 35.6, 36.35, 37.09, 37.86, 38.68, 39.51, 40.34, 41.25] 2020-02-01 18:53:29.053: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:53:29.054: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.15 2020-02-01 18:53:29.054: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 18:53:29.054: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 18:53:29.189: INFO @evaluate_confidence: Previous accuracy would be: 51.58 2020-02-01 18:53:29.189: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 18:53:29.205: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [61.18, 61.78, 62.45, 63.18, 63.92, 64.68, 65.44, 66.29] 2020-02-01 18:53:29.205: INFO @evaluate_confidence: Dropped ratios are: [44.68, 47.63, 50.54, 53.43, 56.26, 58.93, 61.62, 64.19] 2020-02-01 18:53:29.213: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:53:29.213: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.22 2020-02-01 18:53:29.213: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.20 2020-02-01 18:53:29.213: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 18:53:29.308: INFO @evaluate_confidence: Previous accuracy would be: 89.33 2020-02-01 18:53:29.308: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 18:53:29.316: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.33, 94.49, 94.69, 94.93, 95.09, 95.26, 95.42, 95.54, 95.72, 95.88, 96.07, 96.19, 96.31, 96.45, 96.6, 96.75, 96.93, 97.03, 97.14, 97.25, 97.32, 97.4, 97.55, 97.64, 97.77, 97.86, 97.97, 98.03, 98.09, 98.15, 98.25, 98.33, 98.42, 98.56, 98.64] 2020-02-01 18:53:29.316: INFO @evaluate_confidence: Dropped ratios are: [16.17, 16.81, 17.48, 18.29, 18.91, 19.56, 20.2, 20.85, 21.48, 22.1, 22.83, 23.44, 24.15, 24.84, 25.61, 26.37, 27.1, 27.75, 28.52, 29.26, 29.85, 30.51, 31.32, 32.16, 32.88, 33.53, 34.41, 35.14, 35.97, 36.71, 37.6, 38.53, 39.41, 40.32, 41.32] 2020-02-01 18:53:29.323: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:53:29.323: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.12 2020-02-01 18:53:29.323: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 18:53:29.324: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 18:53:29.447: INFO @evaluate_confidence: Previous accuracy would be: 51.95 2020-02-01 18:53:29.447: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 18:53:29.449: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.37, 58.41, 58.78, 58.81, 58.92, 58.84] 2020-02-01 18:53:29.449: INFO @evaluate_confidence: Dropped ratios are: [46.33, 50.07, 53.84, 57.45, 60.88, 64.35] 2020-02-01 18:53:29.502: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:53:30.199: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:53:30.283: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:53:30.732: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:53:30.805: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:53:31.480: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:53:31.565: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:53:31.567: INFO @evaluate_gender-age_model: groups 0 4.865492 1 5.274876 2 6.145460 3 6.147161 4 7.099462 5 7.215486 6 7.419835 7 8.806420 Name: errors, dtype: float64 2020-02-01 18:53:31.568: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:53:32.019: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:53:32.080: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 18:53:32.081: INFO @evaluate_gender-age_model: groups 0 6.980048 1 5.634796 2 5.803451 3 5.879749 4 7.837123 5 5.654971 6 7.426864 7 12.147773 Name: errors, dtype: float64 2020-02-01 18:53:32.241: INFO @decay_lr : LR updated to `7.783126e-05` 2020-02-01 18:53:32.243: INFO @log_profile : T train: 121.844343 2020-02-01 18:53:32.243: INFO @log_profile : T valid: 5.411470 2020-02-01 18:53:32.243: INFO @log_profile : T read data: 1.877863 2020-02-01 18:53:32.243: INFO @log_profile : T hooks: 10.172549 2020-02-01 18:53:32.243: INFO @main_loop : Epoch 50 done 2020-02-01 18:53:32.243: INFO @main_loop : Training epoch 51 2020-02-01 18:55:42.964: INFO @log_variables: train loss nanmean: 0.844803 2020-02-01 18:55:42.965: INFO @log_variables: train age_loss mean: 6.077012 2020-02-01 18:55:42.965: INFO @log_variables: train gender_loss mean: 0.182163 2020-02-01 18:55:42.965: INFO @log_variables: train age_mae mean: 6.555816 2020-02-01 18:55:42.965: INFO @log_variables: train gender_accuracy mean: 0.924745 2020-02-01 18:55:42.965: INFO @log_variables: train gender_confidence/loss nanmean: 0.060427 2020-02-01 18:55:42.965: INFO @log_variables: train gender_confidence/accuracy mean: 0.820072 2020-02-01 18:55:42.965: INFO @log_variables: train age_confidence/loss mean: 0.066317 2020-02-01 18:55:42.965: INFO @log_variables: train age_confidence/accuracy mean: 0.608089 2020-02-01 18:55:42.965: INFO @log_variables: valid loss nanmean: 0.881334 2020-02-01 18:55:42.965: INFO @log_variables: valid age_loss mean: 6.101776 2020-02-01 18:55:42.965: INFO @log_variables: valid gender_loss mean: 0.222349 2020-02-01 18:55:42.965: INFO @log_variables: valid age_mae mean: 6.582303 2020-02-01 18:55:42.965: INFO @log_variables: valid gender_accuracy mean: 0.905493 2020-02-01 18:55:42.965: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056185 2020-02-01 18:55:42.965: INFO @log_variables: valid gender_confidence/accuracy mean: 0.855774 2020-02-01 18:55:42.965: INFO @log_variables: valid age_confidence/loss mean: 0.068307 2020-02-01 18:55:42.965: INFO @log_variables: valid age_confidence/accuracy mean: 0.573105 2020-02-01 18:55:42.966: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:55:42.973: INFO @metrics_hook: train age_mae: 6.556 +-0.037 (110372) 2020-02-01 18:55:42.980: INFO @metrics_hook: train gender_accuracy: 0.925 +-0.002 (110372) 2020-02-01 18:55:45.722: INFO @metrics_hook: valid age_mae: 6.582 +-0.093 (17639) 2020-02-01 18:55:45.723: INFO @metrics_hook: valid gender_accuracy: 0.905 +-0.004 (17639) 2020-02-01 18:55:47.339: INFO @decay_lr : LR updated to `7.7442106e-05` 2020-02-01 18:55:47.340: INFO @log_profile : T train: 121.759060 2020-02-01 18:55:47.340: INFO @log_profile : T valid: 5.476128 2020-02-01 18:55:47.340: INFO @log_profile : T read data: 2.803371 2020-02-01 18:55:47.340: INFO @log_profile : T hooks: 4.982654 2020-02-01 18:55:47.340: INFO @main_loop : Epoch 51 done 2020-02-01 18:55:47.340: INFO @main_loop : Training epoch 52 2020-02-01 18:57:58.009: INFO @log_variables: train loss nanmean: 0.844398 2020-02-01 18:57:58.009: INFO @log_variables: train age_loss mean: 6.102276 2020-02-01 18:57:58.010: INFO @log_variables: train gender_loss mean: 0.179359 2020-02-01 18:57:58.010: INFO @log_variables: train age_mae mean: 6.581443 2020-02-01 18:57:58.010: INFO @log_variables: train gender_accuracy mean: 0.925778 2020-02-01 18:57:58.010: INFO @log_variables: train gender_confidence/loss nanmean: 0.060402 2020-02-01 18:57:58.010: INFO @log_variables: train gender_confidence/accuracy mean: 0.821957 2020-02-01 18:57:58.010: INFO @log_variables: train age_confidence/loss mean: 0.066190 2020-02-01 18:57:58.010: INFO @log_variables: train age_confidence/accuracy mean: 0.607174 2020-02-01 18:57:58.010: INFO @log_variables: valid loss nanmean: 0.858611 2020-02-01 18:57:58.010: INFO @log_variables: valid age_loss mean: 5.924197 2020-02-01 18:57:58.010: INFO @log_variables: valid gender_loss mean: 0.213562 2020-02-01 18:57:58.010: INFO @log_variables: valid age_mae mean: 6.403471 2020-02-01 18:57:58.010: INFO @log_variables: valid gender_accuracy mean: 0.909008 2020-02-01 18:57:58.010: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055897 2020-02-01 18:57:58.010: INFO @log_variables: valid gender_confidence/accuracy mean: 0.849368 2020-02-01 18:57:58.010: INFO @log_variables: valid age_confidence/loss mean: 0.070004 2020-02-01 18:57:58.010: INFO @log_variables: valid age_confidence/accuracy mean: 0.554680 2020-02-01 18:57:58.010: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 18:57:58.017: INFO @metrics_hook: train age_mae: 6.581 +-0.037 (110372) 2020-02-01 18:57:58.024: INFO @metrics_hook: train gender_accuracy: 0.926 +-0.002 (110372) 2020-02-01 18:58:00.788: INFO @metrics_hook: valid age_mae: 6.403 +-0.092 (17639) 2020-02-01 18:58:00.789: INFO @metrics_hook: valid gender_accuracy: 0.909 +-0.004 (17639) 2020-02-01 18:58:02.251: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:58:02.251: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 18:58:02.251: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.20 2020-02-01 18:58:02.251: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 18:58:02.381: INFO @evaluate_confidence: Previous accuracy would be: 92.58 2020-02-01 18:58:02.381: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 18:58:02.446: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.5, 96.61, 96.73, 96.84, 96.96, 97.07, 97.2, 97.29, 97.37, 97.47, 97.54, 97.64, 97.73, 97.8, 97.88, 97.96, 98.02, 98.09, 98.15, 98.21, 98.29, 98.34, 98.41, 98.46, 98.52, 98.58, 98.64, 98.69, 98.74, 98.78, 98.82, 98.86, 98.91, 98.97, 99.01, 99.05, 99.1, 99.14, 99.18] 2020-02-01 18:58:02.446: INFO @evaluate_confidence: Dropped ratios are: [15.58, 16.23, 16.86, 17.49, 18.1, 18.74, 19.41, 20.04, 20.63, 21.24, 21.81, 22.41, 23.03, 23.67, 24.27, 24.92, 25.48, 26.12, 26.71, 27.35, 28.0, 28.62, 29.3, 29.98, 30.61, 31.3, 31.99, 32.68, 33.37, 34.09, 34.8, 35.51, 36.3, 37.01, 37.78, 38.59, 39.38, 40.23, 41.12] 2020-02-01 18:58:02.496: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:58:02.497: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 18:58:02.497: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.12 2020-02-01 18:58:02.497: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 18:58:02.633: INFO @evaluate_confidence: Previous accuracy would be: 51.52 2020-02-01 18:58:02.633: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 18:58:02.654: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [60.48, 61.1, 61.73, 62.43, 63.12, 63.82, 64.62, 65.35, 66.19, 66.96] 2020-02-01 18:58:02.654: INFO @evaluate_confidence: Dropped ratios are: [41.93, 44.75, 47.66, 50.59, 53.39, 56.14, 58.94, 61.46, 63.9, 66.29] 2020-02-01 18:58:02.662: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 18:58:02.662: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.22 2020-02-01 18:58:02.662: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.19 2020-02-01 18:58:02.662: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 18:58:02.758: INFO @evaluate_confidence: Previous accuracy would be: 90.90 2020-02-01 18:58:02.758: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 18:58:02.767: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.49, 95.7, 95.84, 96.01, 96.13, 96.28, 96.43, 96.55, 96.7, 96.85, 97.01, 97.06, 97.12, 97.23, 97.31, 97.39, 97.51, 97.6, 97.73, 97.84, 97.95, 98.02, 98.11, 98.17, 98.25, 98.3, 98.38, 98.48, 98.53, 98.58, 98.66, 98.72, 98.78, 98.84, 98.94, 99.01, 99.07] 2020-02-01 18:58:02.767: INFO @evaluate_confidence: Dropped ratios are: [15.1, 15.72, 16.32, 16.93, 17.48, 18.06, 18.63, 19.39, 20.01, 20.71, 21.22, 21.69, 22.17, 22.77, 23.4, 24.08, 24.7, 25.4, 25.95, 26.65, 27.37, 28.05, 28.81, 29.56, 30.28, 31.1, 31.76, 32.56, 33.34, 34.1, 34.98, 35.95, 36.71, 37.6, 38.61, 39.32, 40.13] 2020-02-01 18:58:02.774: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 18:58:02.775: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 18:58:02.775: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 18:58:02.775: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 18:58:02.901: INFO @evaluate_confidence: Previous accuracy would be: 53.30 2020-02-01 18:58:02.902: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 18:58:02.904: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [60.99, 61.44, 61.46, 61.72, 61.61, 61.83] 2020-02-01 18:58:02.904: INFO @evaluate_confidence: Dropped ratios are: [43.69, 47.65, 51.87, 55.68, 59.58, 62.97] 2020-02-01 18:58:02.958: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 18:58:03.669: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:58:03.755: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 18:58:04.200: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 18:58:04.434: INFO @decay_lr : LR updated to `7.7054894e-05` 2020-02-01 18:58:04.435: INFO @log_profile : T train: 121.680499 2020-02-01 18:58:04.435: INFO @log_profile : T valid: 5.456466 2020-02-01 18:58:04.435: INFO @log_profile : T read data: 2.864029 2020-02-01 18:58:04.435: INFO @log_profile : T hooks: 7.016884 2020-02-01 18:58:04.435: INFO @main_loop : Epoch 52 done 2020-02-01 18:58:04.435: INFO @main_loop : Training epoch 53 2020-02-01 19:00:14.561: INFO @log_variables: train loss nanmean: 0.838731 2020-02-01 19:00:14.561: INFO @log_variables: train age_loss mean: 6.039209 2020-02-01 19:00:14.561: INFO @log_variables: train gender_loss mean: 0.179202 2020-02-01 19:00:14.561: INFO @log_variables: train age_mae mean: 6.518286 2020-02-01 19:00:14.561: INFO @log_variables: train gender_accuracy mean: 0.926016 2020-02-01 19:00:14.561: INFO @log_variables: train gender_confidence/loss nanmean: 0.060241 2020-02-01 19:00:14.561: INFO @log_variables: train gender_confidence/accuracy mean: 0.821551 2020-02-01 19:00:14.561: INFO @log_variables: train age_confidence/loss mean: 0.066561 2020-02-01 19:00:14.562: INFO @log_variables: train age_confidence/accuracy mean: 0.608643 2020-02-01 19:00:14.562: INFO @log_variables: valid loss nanmean: 0.851651 2020-02-01 19:00:14.562: INFO @log_variables: valid age_loss mean: 5.987962 2020-02-01 19:00:14.562: INFO @log_variables: valid gender_loss mean: 0.201904 2020-02-01 19:00:14.562: INFO @log_variables: valid age_mae mean: 6.468578 2020-02-01 19:00:14.562: INFO @log_variables: valid gender_accuracy mean: 0.913771 2020-02-01 19:00:14.562: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054040 2020-02-01 19:00:14.562: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864618 2020-02-01 19:00:14.562: INFO @log_variables: valid age_confidence/loss mean: 0.069702 2020-02-01 19:00:14.562: INFO @log_variables: valid age_confidence/accuracy mean: 0.548104 2020-02-01 19:00:14.562: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:00:14.569: INFO @metrics_hook: train age_mae: 6.518 +-0.037 (110592) 2020-02-01 19:00:14.577: INFO @metrics_hook: train gender_accuracy: 0.926 +-0.002 (110592) 2020-02-01 19:00:17.326: INFO @metrics_hook: valid age_mae: 6.469 +-0.093 (17639) 2020-02-01 19:00:17.328: INFO @metrics_hook: valid gender_accuracy: 0.914 +-0.004 (17639) 2020-02-01 19:00:18.973: INFO @decay_lr : LR updated to `7.666962e-05` 2020-02-01 19:00:19.283: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 19:00:19.285: INFO @log_profile : T train: 121.960899 2020-02-01 19:00:19.286: INFO @log_profile : T valid: 5.560378 2020-02-01 19:00:19.286: INFO @log_profile : T read data: 1.908762 2020-02-01 19:00:19.286: INFO @log_profile : T hooks: 5.342130 2020-02-01 19:00:19.286: INFO @main_loop : Epoch 53 done 2020-02-01 19:00:19.286: INFO @main_loop : Training epoch 54 2020-02-01 19:02:29.949: INFO @log_variables: train loss nanmean: 0.838302 2020-02-01 19:02:29.949: INFO @log_variables: train age_loss mean: 6.037341 2020-02-01 19:02:29.949: INFO @log_variables: train gender_loss mean: 0.178834 2020-02-01 19:02:29.949: INFO @log_variables: train age_mae mean: 6.516233 2020-02-01 19:02:29.949: INFO @log_variables: train gender_accuracy mean: 0.926594 2020-02-01 19:02:29.949: INFO @log_variables: train gender_confidence/loss nanmean: 0.060483 2020-02-01 19:02:29.949: INFO @log_variables: train gender_confidence/accuracy mean: 0.824412 2020-02-01 19:02:29.949: INFO @log_variables: train age_confidence/loss mean: 0.066393 2020-02-01 19:02:29.950: INFO @log_variables: train age_confidence/accuracy mean: 0.609883 2020-02-01 19:02:29.950: INFO @log_variables: valid loss nanmean: 0.858366 2020-02-01 19:02:29.950: INFO @log_variables: valid age_loss mean: 6.021783 2020-02-01 19:02:29.950: INFO @log_variables: valid gender_loss mean: 0.203338 2020-02-01 19:02:29.950: INFO @log_variables: valid age_mae mean: 6.501697 2020-02-01 19:02:29.950: INFO @log_variables: valid gender_accuracy mean: 0.916322 2020-02-01 19:02:29.950: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056784 2020-02-01 19:02:29.950: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868530 2020-02-01 19:02:29.950: INFO @log_variables: valid age_confidence/loss mean: 0.069294 2020-02-01 19:02:29.950: INFO @log_variables: valid age_confidence/accuracy mean: 0.560859 2020-02-01 19:02:29.950: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:02:29.958: INFO @metrics_hook: train age_mae: 6.516 +-0.037 (110372) 2020-02-01 19:02:29.965: INFO @metrics_hook: train gender_accuracy: 0.927 +-0.002 (110372) 2020-02-01 19:02:32.703: INFO @metrics_hook: valid age_mae: 6.502 +-0.093 (17639) 2020-02-01 19:02:32.704: INFO @metrics_hook: valid gender_accuracy: 0.916 +-0.004 (17639) 2020-02-01 19:02:34.156: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:02:34.157: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 19:02:34.157: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.20 2020-02-01 19:02:34.157: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 19:02:34.286: INFO @evaluate_confidence: Previous accuracy would be: 92.66 2020-02-01 19:02:34.286: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 19:02:34.343: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.6, 96.7, 96.81, 96.92, 97.03, 97.13, 97.22, 97.32, 97.41, 97.52, 97.62, 97.72, 97.78, 97.84, 97.91, 97.98, 98.03, 98.1, 98.17, 98.23, 98.3, 98.35, 98.41, 98.47, 98.54, 98.59, 98.65, 98.7, 98.75, 98.82, 98.86, 98.91, 98.95, 98.99, 99.02, 99.05, 99.09, 99.13, 99.16] 2020-02-01 19:02:34.343: INFO @evaluate_confidence: Dropped ratios are: [15.38, 16.0, 16.66, 17.3, 17.93, 18.51, 19.17, 19.77, 20.38, 20.98, 21.58, 22.23, 22.81, 23.45, 24.09, 24.7, 25.29, 25.89, 26.54, 27.15, 27.78, 28.39, 29.08, 29.71, 30.41, 31.13, 31.82, 32.49, 33.21, 33.93, 34.67, 35.43, 36.17, 36.88, 37.66, 38.44, 39.25, 40.11, 41.04] 2020-02-01 19:02:34.392: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:02:34.392: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:02:34.392: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:02:34.393: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 19:02:34.528: INFO @evaluate_confidence: Previous accuracy would be: 51.97 2020-02-01 19:02:34.528: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:02:34.546: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [61.54, 62.26, 62.95, 63.65, 64.29, 65.08, 65.88, 66.61, 67.39] 2020-02-01 19:02:34.546: INFO @evaluate_confidence: Dropped ratios are: [43.79, 46.7, 49.63, 52.54, 55.3, 58.0, 60.65, 63.09, 65.56] 2020-02-01 19:02:34.554: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:02:34.554: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.21 2020-02-01 19:02:34.554: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.20 2020-02-01 19:02:34.554: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.23 2020-02-01 19:02:34.653: INFO @evaluate_confidence: Previous accuracy would be: 91.63 2020-02-01 19:02:34.653: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 19:02:34.661: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.88, 96.05, 96.18, 96.29, 96.42, 96.5, 96.57, 96.7, 96.81, 96.93, 97.02, 97.12, 97.21, 97.27, 97.38, 97.49, 97.57, 97.68, 97.75, 97.87, 97.91, 97.98, 98.14, 98.2, 98.31, 98.4, 98.48, 98.55, 98.62, 98.72, 98.78, 98.86, 98.93, 98.97, 99.04, 99.13] 2020-02-01 19:02:34.661: INFO @evaluate_confidence: Dropped ratios are: [13.94, 14.5, 15.02, 15.55, 15.98, 16.47, 16.88, 17.42, 18.01, 18.5, 19.11, 19.71, 20.24, 20.81, 21.39, 21.99, 22.68, 23.29, 23.93, 24.48, 25.06, 25.7, 26.34, 26.92, 27.54, 28.24, 28.94, 29.68, 30.47, 31.23, 31.95, 32.75, 33.57, 34.43, 35.3, 36.34] 2020-02-01 19:02:34.668: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:02:34.669: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.11 2020-02-01 19:02:34.669: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.10 2020-02-01 19:02:34.669: INFO @evaluate_confidence: Average confidence of all samples 0.46 +- 0.11 2020-02-01 19:02:34.794: INFO @evaluate_confidence: Previous accuracy would be: 52.61 2020-02-01 19:02:34.795: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 19:02:34.796: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [61.77, 62.14, 62.65, 63.03, 63.21, 63.06] 2020-02-01 19:02:34.797: INFO @evaluate_confidence: Dropped ratios are: [43.96, 47.49, 51.51, 55.69, 59.65, 63.98] 2020-02-01 19:02:34.851: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:02:35.573: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:02:35.658: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:02:36.107: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:02:36.338: INFO @decay_lr : LR updated to `7.628627e-05` 2020-02-01 19:02:36.651: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 19:02:36.655: INFO @log_profile : T train: 121.570265 2020-02-01 19:02:36.655: INFO @log_profile : T valid: 5.478036 2020-02-01 19:02:36.655: INFO @log_profile : T read data: 2.914130 2020-02-01 19:02:36.655: INFO @log_profile : T hooks: 7.326951 2020-02-01 19:02:36.655: INFO @main_loop : Epoch 54 done 2020-02-01 19:02:36.655: INFO @main_loop : Training epoch 55 2020-02-01 19:04:47.238: INFO @log_variables: train loss nanmean: 0.832517 2020-02-01 19:04:47.239: INFO @log_variables: train age_loss mean: 6.008037 2020-02-01 19:04:47.239: INFO @log_variables: train gender_loss mean: 0.175982 2020-02-01 19:04:47.239: INFO @log_variables: train age_mae mean: 6.487120 2020-02-01 19:04:47.239: INFO @log_variables: train gender_accuracy mean: 0.927735 2020-02-01 19:04:47.239: INFO @log_variables: train gender_confidence/loss nanmean: 0.059795 2020-02-01 19:04:47.239: INFO @log_variables: train gender_confidence/accuracy mean: 0.824566 2020-02-01 19:04:47.239: INFO @log_variables: train age_confidence/loss mean: 0.066554 2020-02-01 19:04:47.239: INFO @log_variables: train age_confidence/accuracy mean: 0.606186 2020-02-01 19:04:47.239: INFO @log_variables: valid loss nanmean: 0.853498 2020-02-01 19:04:47.239: INFO @log_variables: valid age_loss mean: 5.924924 2020-02-01 19:04:47.239: INFO @log_variables: valid gender_loss mean: 0.208257 2020-02-01 19:04:47.239: INFO @log_variables: valid age_mae mean: 6.404130 2020-02-01 19:04:47.239: INFO @log_variables: valid gender_accuracy mean: 0.912750 2020-02-01 19:04:47.239: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055272 2020-02-01 19:04:47.239: INFO @log_variables: valid gender_confidence/accuracy mean: 0.854017 2020-02-01 19:04:47.239: INFO @log_variables: valid age_confidence/loss mean: 0.070272 2020-02-01 19:04:47.239: INFO @log_variables: valid age_confidence/accuracy mean: 0.555984 2020-02-01 19:04:47.240: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:04:47.247: INFO @metrics_hook: train age_mae: 6.487 +-0.037 (110372) 2020-02-01 19:04:47.253: INFO @metrics_hook: train gender_accuracy: 0.928 +-0.002 (110372) 2020-02-01 19:04:50.026: INFO @metrics_hook: valid age_mae: 6.404 +-0.094 (17639) 2020-02-01 19:04:50.027: INFO @metrics_hook: valid gender_accuracy: 0.913 +-0.004 (17639) 2020-02-01 19:04:51.727: INFO @decay_lr : LR updated to `7.5904834e-05` 2020-02-01 19:04:51.729: INFO @log_profile : T train: 121.649978 2020-02-01 19:04:51.729: INFO @log_profile : T valid: 5.437373 2020-02-01 19:04:51.729: INFO @log_profile : T read data: 2.835555 2020-02-01 19:04:51.729: INFO @log_profile : T hooks: 5.073866 2020-02-01 19:04:51.729: INFO @main_loop : Epoch 55 done 2020-02-01 19:04:51.729: INFO @main_loop : Training epoch 56 2020-02-01 19:07:01.460: INFO @log_variables: train loss nanmean: 0.830701 2020-02-01 19:07:01.460: INFO @log_variables: train age_loss mean: 5.967620 2020-02-01 19:07:01.460: INFO @log_variables: train gender_loss mean: 0.177189 2020-02-01 19:07:01.460: INFO @log_variables: train age_mae mean: 6.446325 2020-02-01 19:07:01.461: INFO @log_variables: train gender_accuracy mean: 0.927201 2020-02-01 19:07:01.461: INFO @log_variables: train gender_confidence/loss nanmean: 0.060350 2020-02-01 19:07:01.461: INFO @log_variables: train gender_confidence/accuracy mean: 0.823812 2020-02-01 19:07:01.461: INFO @log_variables: train age_confidence/loss mean: 0.066760 2020-02-01 19:07:01.461: INFO @log_variables: train age_confidence/accuracy mean: 0.605885 2020-02-01 19:07:01.461: INFO @log_variables: valid loss nanmean: 0.860941 2020-02-01 19:07:01.461: INFO @log_variables: valid age_loss mean: 6.079464 2020-02-01 19:07:01.461: INFO @log_variables: valid gender_loss mean: 0.203921 2020-02-01 19:07:01.461: INFO @log_variables: valid age_mae mean: 6.560685 2020-02-01 19:07:01.461: INFO @log_variables: valid gender_accuracy mean: 0.915528 2020-02-01 19:07:01.461: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053901 2020-02-01 19:07:01.461: INFO @log_variables: valid gender_confidence/accuracy mean: 0.850729 2020-02-01 19:07:01.461: INFO @log_variables: valid age_confidence/loss mean: 0.068979 2020-02-01 19:07:01.461: INFO @log_variables: valid age_confidence/accuracy mean: 0.553943 2020-02-01 19:07:01.461: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:07:01.468: INFO @metrics_hook: train age_mae: 6.446 +-0.037 (110592) 2020-02-01 19:07:01.475: INFO @metrics_hook: train gender_accuracy: 0.927 +-0.002 (110592) 2020-02-01 19:07:04.234: INFO @metrics_hook: valid age_mae: 6.561 +-0.092 (17639) 2020-02-01 19:07:04.236: INFO @metrics_hook: valid gender_accuracy: 0.916 +-0.004 (17639) 2020-02-01 19:07:09.079: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:07:09.079: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 19:07:09.079: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.20 2020-02-01 19:07:09.080: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 19:07:09.212: INFO @evaluate_confidence: Previous accuracy would be: 92.72 2020-02-01 19:07:09.212: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 19:07:09.272: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.58, 96.69, 96.81, 96.91, 97.02, 97.13, 97.22, 97.32, 97.43, 97.52, 97.59, 97.68, 97.77, 97.85, 97.93, 98.02, 98.09, 98.15, 98.21, 98.27, 98.33, 98.4, 98.47, 98.53, 98.58, 98.64, 98.69, 98.74, 98.79, 98.85, 98.88, 98.94, 98.98, 99.02, 99.06, 99.1, 99.13, 99.17, 99.23] 2020-02-01 19:07:09.272: INFO @evaluate_confidence: Dropped ratios are: [15.49, 16.13, 16.73, 17.35, 17.97, 18.57, 19.14, 19.75, 20.37, 20.98, 21.6, 22.21, 22.84, 23.46, 24.07, 24.67, 25.29, 25.91, 26.53, 27.12, 27.73, 28.34, 28.98, 29.63, 30.25, 30.9, 31.55, 32.21, 32.93, 33.63, 34.33, 35.1, 35.86, 36.61, 37.4, 38.17, 38.95, 39.81, 40.76] 2020-02-01 19:07:09.323: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:07:09.323: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:07:09.323: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:07:09.324: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 19:07:09.465: INFO @evaluate_confidence: Previous accuracy would be: 52.20 2020-02-01 19:07:09.465: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:07:09.484: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [61.3, 61.97, 62.65, 63.36, 64.05, 64.81, 65.59, 66.38, 67.15] 2020-02-01 19:07:09.484: INFO @evaluate_confidence: Dropped ratios are: [43.28, 46.24, 49.17, 52.13, 55.02, 57.8, 60.49, 63.13, 65.55] 2020-02-01 19:07:09.492: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:07:09.492: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.22 2020-02-01 19:07:09.493: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.20 2020-02-01 19:07:09.493: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.25 2020-02-01 19:07:09.591: INFO @evaluate_confidence: Previous accuracy would be: 91.55 2020-02-01 19:07:09.592: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 19:07:09.600: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.8, 95.92, 96.07, 96.19, 96.38, 96.49, 96.65, 96.72, 96.78, 96.88, 97.05, 97.12, 97.23, 97.32, 97.41, 97.55, 97.6, 97.68, 97.85, 97.98, 98.07, 98.11, 98.21, 98.27, 98.33, 98.45, 98.52, 98.57, 98.65, 98.68, 98.79, 98.83, 98.9, 98.97, 98.99, 99.07, 99.16, 99.23] 2020-02-01 19:07:09.600: INFO @evaluate_confidence: Dropped ratios are: [14.51, 15.06, 15.6, 16.14, 16.71, 17.33, 17.88, 18.39, 18.95, 19.54, 20.13, 20.68, 21.27, 21.88, 22.46, 23.09, 23.68, 24.33, 24.94, 25.56, 26.14, 26.76, 27.48, 28.06, 28.78, 29.41, 30.12, 30.83, 31.6, 32.33, 33.13, 33.88, 34.65, 35.44, 36.31, 37.21, 38.23, 39.14] 2020-02-01 19:07:09.608: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:07:09.608: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.12 2020-02-01 19:07:09.608: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 19:07:09.609: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 19:07:09.736: INFO @evaluate_confidence: Previous accuracy would be: 51.06 2020-02-01 19:07:09.737: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 19:07:09.738: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.29, 57.52, 57.93, 58.08, 58.58] 2020-02-01 19:07:09.738: INFO @evaluate_confidence: Dropped ratios are: [46.02, 50.11, 54.08, 57.9, 61.66] 2020-02-01 19:07:09.790: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:07:10.490: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:07:10.579: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:07:11.030: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:07:11.267: INFO @decay_lr : LR updated to `7.552531e-05` 2020-02-01 19:07:11.269: INFO @log_profile : T train: 121.665726 2020-02-01 19:07:11.269: INFO @log_profile : T valid: 5.455492 2020-02-01 19:07:11.269: INFO @log_profile : T read data: 1.907656 2020-02-01 19:07:11.269: INFO @log_profile : T hooks: 10.432704 2020-02-01 19:07:11.269: INFO @main_loop : Epoch 56 done 2020-02-01 19:07:11.269: INFO @main_loop : Training epoch 57 2020-02-01 19:09:21.943: INFO @log_variables: train loss nanmean: 0.823834 2020-02-01 19:09:21.943: INFO @log_variables: train age_loss mean: 5.939348 2020-02-01 19:09:21.943: INFO @log_variables: train gender_loss mean: 0.172866 2020-02-01 19:09:21.943: INFO @log_variables: train age_mae mean: 6.418417 2020-02-01 19:09:21.943: INFO @log_variables: train gender_accuracy mean: 0.929620 2020-02-01 19:09:21.943: INFO @log_variables: train gender_confidence/loss nanmean: 0.059718 2020-02-01 19:09:21.943: INFO @log_variables: train gender_confidence/accuracy mean: 0.826659 2020-02-01 19:09:21.943: INFO @log_variables: train age_confidence/loss mean: 0.067025 2020-02-01 19:09:21.944: INFO @log_variables: train age_confidence/accuracy mean: 0.605199 2020-02-01 19:09:21.944: INFO @log_variables: valid loss nanmean: 0.858898 2020-02-01 19:09:21.944: INFO @log_variables: valid age_loss mean: 5.949850 2020-02-01 19:09:21.944: INFO @log_variables: valid gender_loss mean: 0.213719 2020-02-01 19:09:21.944: INFO @log_variables: valid age_mae mean: 6.431153 2020-02-01 19:09:21.944: INFO @log_variables: valid gender_accuracy mean: 0.909916 2020-02-01 19:09:21.944: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053580 2020-02-01 19:09:21.944: INFO @log_variables: valid gender_confidence/accuracy mean: 0.861670 2020-02-01 19:09:21.944: INFO @log_variables: valid age_confidence/loss mean: 0.070133 2020-02-01 19:09:21.944: INFO @log_variables: valid age_confidence/accuracy mean: 0.536935 2020-02-01 19:09:21.944: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:09:21.951: INFO @metrics_hook: train age_mae: 6.418 +-0.037 (110372) 2020-02-01 19:09:21.958: INFO @metrics_hook: train gender_accuracy: 0.930 +-0.002 (110372) 2020-02-01 19:09:24.707: INFO @metrics_hook: valid age_mae: 6.431 +-0.091 (17639) 2020-02-01 19:09:24.708: INFO @metrics_hook: valid gender_accuracy: 0.910 +-0.004 (17639) 2020-02-01 19:09:26.369: INFO @decay_lr : LR updated to `7.5147684e-05` 2020-02-01 19:09:26.370: INFO @log_profile : T train: 121.679833 2020-02-01 19:09:26.370: INFO @log_profile : T valid: 5.515977 2020-02-01 19:09:26.370: INFO @log_profile : T read data: 2.785313 2020-02-01 19:09:26.370: INFO @log_profile : T hooks: 5.042416 2020-02-01 19:09:26.370: INFO @main_loop : Epoch 57 done 2020-02-01 19:09:26.370: INFO @main_loop : Training epoch 58 2020-02-01 19:11:36.919: INFO @log_variables: train loss nanmean: 0.822209 2020-02-01 19:11:36.919: INFO @log_variables: train age_loss mean: 5.901093 2020-02-01 19:11:36.919: INFO @log_variables: train gender_loss mean: 0.174717 2020-02-01 19:11:36.919: INFO @log_variables: train age_mae mean: 6.379587 2020-02-01 19:11:36.919: INFO @log_variables: train gender_accuracy mean: 0.928705 2020-02-01 19:11:36.919: INFO @log_variables: train gender_confidence/loss nanmean: 0.059999 2020-02-01 19:11:36.919: INFO @log_variables: train gender_confidence/accuracy mean: 0.825689 2020-02-01 19:11:36.919: INFO @log_variables: train age_confidence/loss mean: 0.066912 2020-02-01 19:11:36.919: INFO @log_variables: train age_confidence/accuracy mean: 0.611378 2020-02-01 19:11:36.919: INFO @log_variables: valid loss nanmean: 0.855064 2020-02-01 19:11:36.919: INFO @log_variables: valid age_loss mean: 5.995118 2020-02-01 19:11:36.920: INFO @log_variables: valid gender_loss mean: 0.204595 2020-02-01 19:11:36.920: INFO @log_variables: valid age_mae mean: 6.476932 2020-02-01 19:11:36.920: INFO @log_variables: valid gender_accuracy mean: 0.913147 2020-02-01 19:11:36.920: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054612 2020-02-01 19:11:36.920: INFO @log_variables: valid gender_confidence/accuracy mean: 0.851352 2020-02-01 19:11:36.920: INFO @log_variables: valid age_confidence/loss mean: 0.069446 2020-02-01 19:11:36.920: INFO @log_variables: valid age_confidence/accuracy mean: 0.554113 2020-02-01 19:11:36.920: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:11:36.927: INFO @metrics_hook: train age_mae: 6.380 +-0.036 (110372) 2020-02-01 19:11:36.934: INFO @metrics_hook: train gender_accuracy: 0.929 +-0.002 (110372) 2020-02-01 19:11:39.770: INFO @metrics_hook: valid age_mae: 6.477 +-0.091 (17639) 2020-02-01 19:11:39.772: INFO @metrics_hook: valid gender_accuracy: 0.913 +-0.004 (17639) 2020-02-01 19:11:41.242: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:11:41.242: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 19:11:41.242: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.41 +- 0.20 2020-02-01 19:11:41.242: INFO @evaluate_confidence: Average confidence of all samples 0.76 +- 0.26 2020-02-01 19:11:41.370: INFO @evaluate_confidence: Previous accuracy would be: 92.87 2020-02-01 19:11:41.371: INFO @evaluate_confidence: Possible optimal thresholds are: [0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 19:11:41.428: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.7, 96.85, 96.96, 97.08, 97.16, 97.25, 97.34, 97.4, 97.48, 97.55, 97.63, 97.71, 97.79, 97.87, 97.97, 98.04, 98.12, 98.19, 98.25, 98.31, 98.39, 98.43, 98.48, 98.54, 98.61, 98.67, 98.72, 98.77, 98.82, 98.86, 98.92, 98.95, 99.0, 99.04, 99.08, 99.13, 99.16, 99.19, 99.23] 2020-02-01 19:11:41.429: INFO @evaluate_confidence: Dropped ratios are: [15.36, 16.0, 16.63, 17.18, 17.75, 18.36, 18.97, 19.54, 20.12, 20.68, 21.22, 21.8, 22.4, 23.0, 23.63, 24.2, 24.8, 25.41, 26.03, 26.61, 27.23, 27.88, 28.52, 29.19, 29.83, 30.52, 31.15, 31.82, 32.5, 33.24, 33.99, 34.73, 35.48, 36.27, 37.08, 37.9, 38.71, 39.53, 40.37] 2020-02-01 19:11:41.481: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:11:41.481: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:11:41.481: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:11:41.481: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 19:11:41.617: INFO @evaluate_confidence: Previous accuracy would be: 52.78 2020-02-01 19:11:41.617: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:11:41.636: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [62.28, 62.95, 63.65, 64.31, 65.07, 65.74, 66.52, 67.22, 68.07] 2020-02-01 19:11:41.636: INFO @evaluate_confidence: Dropped ratios are: [42.72, 45.62, 48.55, 51.39, 54.22, 57.08, 59.79, 62.3, 64.85] 2020-02-01 19:11:41.644: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:11:41.644: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.22 2020-02-01 19:11:41.644: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.20 2020-02-01 19:11:41.644: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 19:11:41.744: INFO @evaluate_confidence: Previous accuracy would be: 91.31 2020-02-01 19:11:41.744: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 19:11:41.753: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.82, 95.96, 96.13, 96.24, 96.38, 96.51, 96.65, 96.72, 96.88, 96.95, 97.07, 97.17, 97.28, 97.41, 97.5, 97.57, 97.6, 97.68, 97.78, 97.89, 97.96, 98.03, 98.11, 98.17, 98.23, 98.31, 98.37, 98.4, 98.46, 98.5, 98.6, 98.64, 98.7, 98.78, 98.87, 98.92, 98.99, 99.04] 2020-02-01 19:11:41.753: INFO @evaluate_confidence: Dropped ratios are: [14.6, 15.17, 15.81, 16.38, 16.96, 17.5, 18.06, 18.67, 19.25, 19.73, 20.2, 20.73, 21.21, 21.77, 22.37, 22.93, 23.54, 24.11, 24.73, 25.43, 26.0, 26.58, 27.22, 27.92, 28.55, 29.3, 30.02, 30.7, 31.46, 32.12, 32.82, 33.52, 34.28, 35.1, 36.11, 36.93, 37.79, 38.71] 2020-02-01 19:11:41.760: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:11:41.761: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 19:11:41.761: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 19:11:41.761: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 19:11:41.890: INFO @evaluate_confidence: Previous accuracy would be: 52.04 2020-02-01 19:11:41.890: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 19:11:41.892: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.07, 59.27, 59.48, 59.75] 2020-02-01 19:11:41.892: INFO @evaluate_confidence: Dropped ratios are: [47.89, 51.54, 55.19, 58.76] 2020-02-01 19:11:41.944: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:11:42.646: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:11:42.732: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:11:43.191: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:11:43.435: INFO @decay_lr : LR updated to `7.4771946e-05` 2020-02-01 19:11:43.436: INFO @log_profile : T train: 121.570025 2020-02-01 19:11:43.436: INFO @log_profile : T valid: 5.443367 2020-02-01 19:11:43.436: INFO @log_profile : T read data: 2.832990 2020-02-01 19:11:43.436: INFO @log_profile : T hooks: 7.140969 2020-02-01 19:11:43.436: INFO @main_loop : Epoch 58 done 2020-02-01 19:11:43.436: INFO @main_loop : Training epoch 59 2020-02-01 19:13:53.383: INFO @log_variables: train loss nanmean: 0.817218 2020-02-01 19:13:53.383: INFO @log_variables: train age_loss mean: 5.913505 2020-02-01 19:13:53.383: INFO @log_variables: train gender_loss mean: 0.169081 2020-02-01 19:13:53.383: INFO @log_variables: train age_mae mean: 6.392311 2020-02-01 19:13:53.383: INFO @log_variables: train gender_accuracy mean: 0.931532 2020-02-01 19:13:53.383: INFO @log_variables: train gender_confidence/loss nanmean: 0.058980 2020-02-01 19:13:53.384: INFO @log_variables: train gender_confidence/accuracy mean: 0.828740 2020-02-01 19:13:53.384: INFO @log_variables: train age_confidence/loss mean: 0.066936 2020-02-01 19:13:53.384: INFO @log_variables: train age_confidence/accuracy mean: 0.606292 2020-02-01 19:13:53.384: INFO @log_variables: valid loss nanmean: 0.860944 2020-02-01 19:13:53.384: INFO @log_variables: valid age_loss mean: 5.953249 2020-02-01 19:13:53.384: INFO @log_variables: valid gender_loss mean: 0.214408 2020-02-01 19:13:53.384: INFO @log_variables: valid age_mae mean: 6.433625 2020-02-01 19:13:53.384: INFO @log_variables: valid gender_accuracy mean: 0.910709 2020-02-01 19:13:53.384: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055097 2020-02-01 19:13:53.384: INFO @log_variables: valid gender_confidence/accuracy mean: 0.859913 2020-02-01 19:13:53.384: INFO @log_variables: valid age_confidence/loss mean: 0.069727 2020-02-01 19:13:53.384: INFO @log_variables: valid age_confidence/accuracy mean: 0.551392 2020-02-01 19:13:53.384: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:13:53.391: INFO @metrics_hook: train age_mae: 6.392 +-0.036 (110592) 2020-02-01 19:13:53.398: INFO @metrics_hook: train gender_accuracy: 0.932 +-0.002 (110592) 2020-02-01 19:13:56.077: INFO @metrics_hook: valid age_mae: 6.434 +-0.090 (17639) 2020-02-01 19:13:56.078: INFO @metrics_hook: valid gender_accuracy: 0.911 +-0.004 (17639) 2020-02-01 19:13:57.716: INFO @decay_lr : LR updated to `7.4398085e-05` 2020-02-01 19:13:57.717: INFO @log_profile : T train: 121.874777 2020-02-01 19:13:57.717: INFO @log_profile : T valid: 5.548272 2020-02-01 19:13:57.717: INFO @log_profile : T read data: 1.860299 2020-02-01 19:13:57.717: INFO @log_profile : T hooks: 4.921047 2020-02-01 19:13:57.717: INFO @main_loop : Epoch 59 done 2020-02-01 19:13:57.718: INFO @main_loop : Training epoch 60 2020-02-01 19:16:09.992: INFO @log_variables: train loss nanmean: 0.814542 2020-02-01 19:16:09.993: INFO @log_variables: train age_loss mean: 5.875745 2020-02-01 19:16:09.993: INFO @log_variables: train gender_loss mean: 0.169308 2020-02-01 19:16:09.993: INFO @log_variables: train age_mae mean: 6.354942 2020-02-01 19:16:09.993: INFO @log_variables: train gender_accuracy mean: 0.931405 2020-02-01 19:16:09.993: INFO @log_variables: train gender_confidence/loss nanmean: 0.059357 2020-02-01 19:16:09.993: INFO @log_variables: train gender_confidence/accuracy mean: 0.828453 2020-02-01 19:16:09.993: INFO @log_variables: train age_confidence/loss mean: 0.067111 2020-02-01 19:16:09.993: INFO @log_variables: train age_confidence/accuracy mean: 0.608470 2020-02-01 19:16:09.993: INFO @log_variables: valid loss nanmean: 0.875811 2020-02-01 19:16:09.993: INFO @log_variables: valid age_loss mean: 5.974945 2020-02-01 19:16:09.993: INFO @log_variables: valid gender_loss mean: 0.228307 2020-02-01 19:16:09.993: INFO @log_variables: valid age_mae mean: 6.454862 2020-02-01 19:16:09.993: INFO @log_variables: valid gender_accuracy mean: 0.904983 2020-02-01 19:16:09.993: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055186 2020-02-01 19:16:09.993: INFO @log_variables: valid gender_confidence/accuracy mean: 0.853166 2020-02-01 19:16:09.993: INFO @log_variables: valid age_confidence/loss mean: 0.069897 2020-02-01 19:16:09.993: INFO @log_variables: valid age_confidence/accuracy mean: 0.550371 2020-02-01 19:16:09.993: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:16:10.001: INFO @metrics_hook: train age_mae: 6.355 +-0.036 (110372) 2020-02-01 19:16:10.008: INFO @metrics_hook: train gender_accuracy: 0.931 +-0.002 (110372) 2020-02-01 19:16:12.732: INFO @metrics_hook: valid age_mae: 6.455 +-0.092 (17639) 2020-02-01 19:16:12.733: INFO @metrics_hook: valid gender_accuracy: 0.905 +-0.004 (17639) 2020-02-01 19:16:14.182: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:16:14.183: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 19:16:14.183: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.20 2020-02-01 19:16:14.183: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:16:14.314: INFO @evaluate_confidence: Previous accuracy would be: 93.14 2020-02-01 19:16:14.315: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 19:16:14.374: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.8, 96.91, 97.01, 97.12, 97.21, 97.31, 97.39, 97.5, 97.57, 97.65, 97.73, 97.8, 97.88, 97.95, 98.02, 98.1, 98.15, 98.21, 98.27, 98.32, 98.37, 98.43, 98.49, 98.53, 98.61, 98.66, 98.71, 98.75, 98.8, 98.84, 98.89, 98.93, 98.98, 99.02, 99.06, 99.12, 99.16, 99.2, 99.24, 99.27] 2020-02-01 19:16:14.374: INFO @evaluate_confidence: Dropped ratios are: [14.59, 15.18, 15.76, 16.4, 17.01, 17.59, 18.15, 18.7, 19.27, 19.83, 20.4, 20.95, 21.53, 22.09, 22.65, 23.24, 23.8, 24.42, 25.01, 25.62, 26.19, 26.79, 27.43, 28.01, 28.69, 29.33, 29.98, 30.66, 31.33, 32.04, 32.72, 33.43, 34.12, 34.84, 35.62, 36.42, 37.23, 38.02, 38.85, 39.69] 2020-02-01 19:16:14.424: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:16:14.425: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:16:14.425: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:16:14.425: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 19:16:14.565: INFO @evaluate_confidence: Previous accuracy would be: 52.93 2020-02-01 19:16:14.565: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:16:14.584: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [62.18, 62.85, 63.47, 64.16, 64.76, 65.49, 66.26, 67.03, 67.76] 2020-02-01 19:16:14.585: INFO @evaluate_confidence: Dropped ratios are: [42.57, 45.53, 48.46, 51.36, 54.21, 57.12, 59.86, 62.53, 64.96] 2020-02-01 19:16:14.593: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:16:14.593: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.22 2020-02-01 19:16:14.593: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.20 2020-02-01 19:16:14.593: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.24 2020-02-01 19:16:14.693: INFO @evaluate_confidence: Previous accuracy would be: 90.50 2020-02-01 19:16:14.694: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 19:16:14.702: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.05, 95.2, 95.38, 95.5, 95.64, 95.76, 95.91, 95.97, 96.12, 96.29, 96.46, 96.59, 96.71, 96.86, 96.94, 97.04, 97.14, 97.28, 97.38, 97.48, 97.63, 97.73, 97.85, 97.9, 98.01, 98.1, 98.19, 98.24, 98.36, 98.45, 98.52, 98.57, 98.62, 98.68, 98.78, 98.88] 2020-02-01 19:16:14.702: INFO @evaluate_confidence: Dropped ratios are: [14.77, 15.34, 15.86, 16.38, 16.95, 17.52, 18.2, 18.7, 19.29, 19.93, 20.47, 21.09, 21.59, 22.23, 22.82, 23.4, 23.98, 24.65, 25.35, 26.07, 26.88, 27.63, 28.26, 28.82, 29.65, 30.35, 31.26, 31.95, 32.77, 33.47, 34.25, 35.08, 35.84, 36.82, 37.98, 38.79] 2020-02-01 19:16:14.710: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:16:14.710: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 19:16:14.710: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 19:16:14.711: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.11 2020-02-01 19:16:14.839: INFO @evaluate_confidence: Previous accuracy would be: 52.36 2020-02-01 19:16:14.839: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 19:16:14.841: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.75, 58.77, 59.02, 59.13] 2020-02-01 19:16:14.841: INFO @evaluate_confidence: Dropped ratios are: [47.67, 51.92, 55.83, 59.51] 2020-02-01 19:16:14.892: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:16:15.589: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:16:15.673: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:16:16.128: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:16:16.206: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:16:16.896: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:16:16.980: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 19:16:16.982: INFO @evaluate_gender-age_model: groups 0 4.518390 1 4.981578 2 5.978079 3 6.062467 4 6.930927 5 6.879633 6 7.074730 7 8.253444 Name: errors, dtype: float64 2020-02-01 19:16:16.983: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:16:17.431: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:16:17.494: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 19:16:17.495: INFO @evaluate_gender-age_model: groups 0 6.534326 1 5.143446 2 5.536380 3 5.669893 4 7.601780 5 5.761235 6 8.089875 7 12.869470 Name: errors, dtype: float64 2020-02-01 19:16:17.663: INFO @decay_lr : LR updated to `7.4026095e-05` 2020-02-01 19:16:17.665: INFO @log_profile : T train: 121.493033 2020-02-01 19:16:17.665: INFO @log_profile : T valid: 5.397606 2020-02-01 19:16:17.665: INFO @log_profile : T read data: 2.813668 2020-02-01 19:16:17.665: INFO @log_profile : T hooks: 10.166558 2020-02-01 19:16:17.665: INFO @main_loop : Epoch 60 done 2020-02-01 19:16:17.665: INFO @main_loop : Training epoch 61 2020-02-01 19:18:28.780: INFO @log_variables: train loss nanmean: 0.812192 2020-02-01 19:18:28.780: INFO @log_variables: train age_loss mean: 5.863907 2020-02-01 19:18:28.781: INFO @log_variables: train gender_loss mean: 0.168333 2020-02-01 19:18:28.781: INFO @log_variables: train age_mae mean: 6.342936 2020-02-01 19:18:28.781: INFO @log_variables: train gender_accuracy mean: 0.930354 2020-02-01 19:18:28.781: INFO @log_variables: train gender_confidence/loss nanmean: 0.058709 2020-02-01 19:18:28.781: INFO @log_variables: train gender_confidence/accuracy mean: 0.830972 2020-02-01 19:18:28.781: INFO @log_variables: train age_confidence/loss mean: 0.067370 2020-02-01 19:18:28.781: INFO @log_variables: train age_confidence/accuracy mean: 0.607727 2020-02-01 19:18:28.781: INFO @log_variables: valid loss nanmean: 0.855848 2020-02-01 19:18:28.781: INFO @log_variables: valid age_loss mean: 5.929244 2020-02-01 19:18:28.781: INFO @log_variables: valid gender_loss mean: 0.211406 2020-02-01 19:18:28.781: INFO @log_variables: valid age_mae mean: 6.409565 2020-02-01 19:18:28.781: INFO @log_variables: valid gender_accuracy mean: 0.910823 2020-02-01 19:18:28.781: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055081 2020-02-01 19:18:28.781: INFO @log_variables: valid gender_confidence/accuracy mean: 0.854300 2020-02-01 19:18:28.781: INFO @log_variables: valid age_confidence/loss mean: 0.069558 2020-02-01 19:18:28.781: INFO @log_variables: valid age_confidence/accuracy mean: 0.565338 2020-02-01 19:18:28.781: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:18:28.788: INFO @metrics_hook: train age_mae: 6.343 +-0.037 (110372) 2020-02-01 19:18:28.795: INFO @metrics_hook: train gender_accuracy: 0.930 +-0.002 (110372) 2020-02-01 19:18:31.561: INFO @metrics_hook: valid age_mae: 6.410 +-0.091 (17639) 2020-02-01 19:18:31.562: INFO @metrics_hook: valid gender_accuracy: 0.911 +-0.004 (17639) 2020-02-01 19:18:33.207: INFO @decay_lr : LR updated to `7.365597e-05` 2020-02-01 19:18:33.208: INFO @log_profile : T train: 121.877816 2020-02-01 19:18:33.208: INFO @log_profile : T valid: 5.554776 2020-02-01 19:18:33.208: INFO @log_profile : T read data: 2.987573 2020-02-01 19:18:33.208: INFO @log_profile : T hooks: 5.045827 2020-02-01 19:18:33.208: INFO @main_loop : Epoch 61 done 2020-02-01 19:18:33.208: INFO @main_loop : Training epoch 62 2020-02-01 19:20:42.903: INFO @log_variables: train loss nanmean: 0.807235 2020-02-01 19:20:42.903: INFO @log_variables: train age_loss mean: 5.825071 2020-02-01 19:20:42.903: INFO @log_variables: train gender_loss mean: 0.166807 2020-02-01 19:20:42.903: INFO @log_variables: train age_mae mean: 6.303042 2020-02-01 19:20:42.903: INFO @log_variables: train gender_accuracy mean: 0.931921 2020-02-01 19:20:42.903: INFO @log_variables: train gender_confidence/loss nanmean: 0.058791 2020-02-01 19:20:42.903: INFO @log_variables: train gender_confidence/accuracy mean: 0.828722 2020-02-01 19:20:42.903: INFO @log_variables: train age_confidence/loss mean: 0.067249 2020-02-01 19:20:42.903: INFO @log_variables: train age_confidence/accuracy mean: 0.606997 2020-02-01 19:20:42.903: INFO @log_variables: valid loss nanmean: 0.869613 2020-02-01 19:20:42.903: INFO @log_variables: valid age_loss mean: 5.937225 2020-02-01 19:20:42.904: INFO @log_variables: valid gender_loss mean: 0.225670 2020-02-01 19:20:42.904: INFO @log_variables: valid age_mae mean: 6.417134 2020-02-01 19:20:42.904: INFO @log_variables: valid gender_accuracy mean: 0.903226 2020-02-01 19:20:42.904: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054385 2020-02-01 19:20:42.904: INFO @log_variables: valid gender_confidence/accuracy mean: 0.850899 2020-02-01 19:20:42.904: INFO @log_variables: valid age_confidence/loss mean: 0.070327 2020-02-01 19:20:42.904: INFO @log_variables: valid age_confidence/accuracy mean: 0.543682 2020-02-01 19:20:42.904: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:20:42.911: INFO @metrics_hook: train age_mae: 6.303 +-0.036 (110592) 2020-02-01 19:20:42.918: INFO @metrics_hook: train gender_accuracy: 0.932 +-0.002 (110592) 2020-02-01 19:20:45.615: INFO @metrics_hook: valid age_mae: 6.417 +-0.093 (17639) 2020-02-01 19:20:45.616: INFO @metrics_hook: valid gender_accuracy: 0.903 +-0.004 (17639) 2020-02-01 19:20:47.062: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:20:47.063: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.79 +- 0.24 2020-02-01 19:20:47.063: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.20 2020-02-01 19:20:47.063: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:20:47.191: INFO @evaluate_confidence: Previous accuracy would be: 93.19 2020-02-01 19:20:47.191: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79] 2020-02-01 19:20:47.250: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.87, 96.98, 97.11, 97.21, 97.31, 97.41, 97.5, 97.57, 97.65, 97.73, 97.8, 97.87, 97.93, 98.01, 98.07, 98.14, 98.21, 98.27, 98.32, 98.39, 98.45, 98.49, 98.54, 98.6, 98.65, 98.69, 98.74, 98.79, 98.82, 98.87, 98.94, 98.99, 99.03, 99.08, 99.1, 99.15, 99.18, 99.21, 99.24, 99.29] 2020-02-01 19:20:47.250: INFO @evaluate_confidence: Dropped ratios are: [14.62, 15.19, 15.8, 16.42, 17.03, 17.65, 18.22, 18.76, 19.31, 19.88, 20.44, 21.01, 21.58, 22.18, 22.73, 23.3, 23.89, 24.45, 24.99, 25.58, 26.15, 26.74, 27.31, 27.91, 28.52, 29.13, 29.75, 30.35, 30.96, 31.64, 32.3, 32.98, 33.75, 34.48, 35.17, 35.96, 36.78, 37.6, 38.4, 39.23] 2020-02-01 19:20:47.298: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:20:47.298: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:20:47.299: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:20:47.299: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.14 2020-02-01 19:20:47.433: INFO @evaluate_confidence: Previous accuracy would be: 53.24 2020-02-01 19:20:47.433: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:20:47.452: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [62.24, 62.79, 63.39, 64.13, 64.88, 65.6, 66.27, 67.04, 67.97] 2020-02-01 19:20:47.452: INFO @evaluate_confidence: Dropped ratios are: [41.82, 44.74, 47.7, 50.69, 53.59, 56.44, 59.19, 61.84, 64.36] 2020-02-01 19:20:47.460: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:20:47.460: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.22 2020-02-01 19:20:47.460: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.20 2020-02-01 19:20:47.460: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.24 2020-02-01 19:20:47.562: INFO @evaluate_confidence: Previous accuracy would be: 90.32 2020-02-01 19:20:47.562: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 19:20:47.571: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [94.92, 95.1, 95.3, 95.51, 95.62, 95.72, 95.87, 96.0, 96.15, 96.24, 96.35, 96.48, 96.61, 96.75, 96.87, 97.01, 97.2, 97.38, 97.48, 97.63, 97.73, 97.77, 97.9, 97.98, 98.08, 98.21, 98.3, 98.37, 98.44, 98.5, 98.62, 98.71, 98.77, 98.81, 98.9, 99.02, 99.12, 99.19] 2020-02-01 19:20:47.571: INFO @evaluate_confidence: Dropped ratios are: [14.7, 15.35, 15.89, 16.49, 16.98, 17.54, 18.06, 18.61, 19.21, 19.85, 20.42, 20.95, 21.54, 22.12, 22.73, 23.33, 23.99, 24.75, 25.45, 26.1, 26.68, 27.15, 27.75, 28.39, 29.13, 29.84, 30.63, 31.33, 32.1, 32.88, 33.85, 34.76, 35.64, 36.59, 37.49, 38.38, 39.38, 40.4] 2020-02-01 19:20:47.578: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:20:47.579: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.12 2020-02-01 19:20:47.579: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 19:20:47.579: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 19:20:47.711: INFO @evaluate_confidence: Previous accuracy would be: 52.69 2020-02-01 19:20:47.711: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 19:20:47.713: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.19, 58.47, 58.73, 59.21] 2020-02-01 19:20:47.713: INFO @evaluate_confidence: Dropped ratios are: [46.82, 50.73, 54.64, 58.35] 2020-02-01 19:20:47.763: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:20:48.454: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:20:48.540: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:20:48.986: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:20:49.230: INFO @decay_lr : LR updated to `7.328769e-05` 2020-02-01 19:20:49.232: INFO @log_profile : T train: 121.670401 2020-02-01 19:20:49.232: INFO @log_profile : T valid: 5.445585 2020-02-01 19:20:49.232: INFO @log_profile : T read data: 1.889616 2020-02-01 19:20:49.232: INFO @log_profile : T hooks: 6.940351 2020-02-01 19:20:49.232: INFO @main_loop : Epoch 62 done 2020-02-01 19:20:49.232: INFO @main_loop : Training epoch 63 2020-02-01 19:23:00.075: INFO @log_variables: train loss nanmean: 0.806444 2020-02-01 19:23:00.075: INFO @log_variables: train age_loss mean: 5.819473 2020-02-01 19:23:00.075: INFO @log_variables: train gender_loss mean: 0.166360 2020-02-01 19:23:00.075: INFO @log_variables: train age_mae mean: 6.298162 2020-02-01 19:23:00.075: INFO @log_variables: train gender_accuracy mean: 0.932564 2020-02-01 19:23:00.075: INFO @log_variables: train gender_confidence/loss nanmean: 0.058684 2020-02-01 19:23:00.075: INFO @log_variables: train gender_confidence/accuracy mean: 0.829132 2020-02-01 19:23:00.075: INFO @log_variables: train age_confidence/loss mean: 0.067482 2020-02-01 19:23:00.075: INFO @log_variables: train age_confidence/accuracy mean: 0.607011 2020-02-01 19:23:00.075: INFO @log_variables: valid loss nanmean: 0.869879 2020-02-01 19:23:00.075: INFO @log_variables: valid age_loss mean: 6.018878 2020-02-01 19:23:00.075: INFO @log_variables: valid gender_loss mean: 0.216719 2020-02-01 19:23:00.075: INFO @log_variables: valid age_mae mean: 6.500644 2020-02-01 19:23:00.075: INFO @log_variables: valid gender_accuracy mean: 0.907308 2020-02-01 19:23:00.075: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055530 2020-02-01 19:23:00.076: INFO @log_variables: valid gender_confidence/accuracy mean: 0.847043 2020-02-01 19:23:00.076: INFO @log_variables: valid age_confidence/loss mean: 0.070161 2020-02-01 19:23:00.076: INFO @log_variables: valid age_confidence/accuracy mean: 0.539940 2020-02-01 19:23:00.076: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:23:00.083: INFO @metrics_hook: train age_mae: 6.298 +-0.036 (110372) 2020-02-01 19:23:00.090: INFO @metrics_hook: train gender_accuracy: 0.933 +-0.002 (110372) 2020-02-01 19:23:02.789: INFO @metrics_hook: valid age_mae: 6.501 +-0.094 (17639) 2020-02-01 19:23:02.790: INFO @metrics_hook: valid gender_accuracy: 0.907 +-0.004 (17639) 2020-02-01 19:23:04.408: INFO @decay_lr : LR updated to `7.292125e-05` 2020-02-01 19:23:04.409: INFO @log_profile : T train: 121.776311 2020-02-01 19:23:04.410: INFO @log_profile : T valid: 5.554987 2020-02-01 19:23:04.410: INFO @log_profile : T read data: 2.817709 2020-02-01 19:23:04.410: INFO @log_profile : T hooks: 4.951636 2020-02-01 19:23:04.410: INFO @main_loop : Epoch 63 done 2020-02-01 19:23:04.410: INFO @main_loop : Training epoch 64 2020-02-01 19:25:14.877: INFO @log_variables: train loss nanmean: 0.803229 2020-02-01 19:25:14.877: INFO @log_variables: train age_loss mean: 5.792785 2020-02-01 19:25:14.877: INFO @log_variables: train gender_loss mean: 0.165654 2020-02-01 19:25:14.878: INFO @log_variables: train age_mae mean: 6.270621 2020-02-01 19:25:14.878: INFO @log_variables: train gender_accuracy mean: 0.932021 2020-02-01 19:25:14.878: INFO @log_variables: train gender_confidence/loss nanmean: 0.058357 2020-02-01 19:25:14.878: INFO @log_variables: train gender_confidence/accuracy mean: 0.830174 2020-02-01 19:25:14.878: INFO @log_variables: train age_confidence/loss mean: 0.067661 2020-02-01 19:25:14.878: INFO @log_variables: train age_confidence/accuracy mean: 0.607138 2020-02-01 19:25:14.878: INFO @log_variables: valid loss nanmean: 0.863696 2020-02-01 19:25:14.878: INFO @log_variables: valid age_loss mean: 5.987313 2020-02-01 19:25:14.878: INFO @log_variables: valid gender_loss mean: 0.214441 2020-02-01 19:25:14.878: INFO @log_variables: valid age_mae mean: 6.467568 2020-02-01 19:25:14.878: INFO @log_variables: valid gender_accuracy mean: 0.910596 2020-02-01 19:25:14.878: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055063 2020-02-01 19:25:14.878: INFO @log_variables: valid gender_confidence/accuracy mean: 0.857078 2020-02-01 19:25:14.878: INFO @log_variables: valid age_confidence/loss mean: 0.069386 2020-02-01 19:25:14.878: INFO @log_variables: valid age_confidence/accuracy mean: 0.551165 2020-02-01 19:25:14.878: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:25:14.885: INFO @metrics_hook: train age_mae: 6.271 +-0.036 (110372) 2020-02-01 19:25:14.892: INFO @metrics_hook: train gender_accuracy: 0.932 +-0.002 (110372) 2020-02-01 19:25:17.617: INFO @metrics_hook: valid age_mae: 6.468 +-0.092 (17639) 2020-02-01 19:25:17.619: INFO @metrics_hook: valid gender_accuracy: 0.911 +-0.004 (17639) 2020-02-01 19:25:19.069: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:25:19.069: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:25:19.069: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.20 2020-02-01 19:25:19.070: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:25:19.196: INFO @evaluate_confidence: Previous accuracy would be: 93.20 2020-02-01 19:25:19.197: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:25:19.257: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.91, 97.02, 97.13, 97.24, 97.33, 97.43, 97.52, 97.59, 97.67, 97.74, 97.82, 97.9, 97.97, 98.05, 98.11, 98.17, 98.23, 98.3, 98.37, 98.45, 98.49, 98.54, 98.61, 98.66, 98.72, 98.77, 98.81, 98.85, 98.89, 98.93, 98.97, 99.0, 99.04, 99.07, 99.11, 99.15, 99.19, 99.23, 99.27, 99.3, 99.34] 2020-02-01 19:25:19.257: INFO @evaluate_confidence: Dropped ratios are: [14.65, 15.23, 15.85, 16.41, 16.96, 17.55, 18.08, 18.65, 19.21, 19.75, 20.3, 20.87, 21.37, 21.96, 22.54, 23.09, 23.66, 24.22, 24.79, 25.38, 25.97, 26.55, 27.15, 27.76, 28.36, 28.99, 29.58, 30.22, 30.9, 31.52, 32.16, 32.79, 33.47, 34.2, 34.93, 35.71, 36.5, 37.29, 38.08, 38.93, 39.82] 2020-02-01 19:25:19.305: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:25:19.306: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.15 2020-02-01 19:25:19.306: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.12 2020-02-01 19:25:19.306: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:25:19.438: INFO @evaluate_confidence: Previous accuracy would be: 53.66 2020-02-01 19:25:19.439: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:25:19.457: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [62.5, 63.12, 63.77, 64.49, 65.23, 65.94, 66.74, 67.41, 68.24] 2020-02-01 19:25:19.458: INFO @evaluate_confidence: Dropped ratios are: [41.45, 44.41, 47.42, 50.4, 53.31, 56.17, 58.9, 61.54, 64.06] 2020-02-01 19:25:19.466: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:25:19.466: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.22 2020-02-01 19:25:19.466: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.20 2020-02-01 19:25:19.466: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.24 2020-02-01 19:25:19.569: INFO @evaluate_confidence: Previous accuracy would be: 91.06 2020-02-01 19:25:19.569: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 19:25:19.578: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.57, 95.73, 95.86, 96.01, 96.17, 96.3, 96.44, 96.59, 96.73, 96.87, 96.99, 97.09, 97.16, 97.25, 97.4, 97.5, 97.62, 97.67, 97.78, 97.85, 97.89, 97.99, 98.08, 98.12, 98.22, 98.25, 98.32, 98.38, 98.43, 98.48, 98.58, 98.65, 98.7, 98.73, 98.79, 98.84, 98.9, 98.98] 2020-02-01 19:25:19.578: INFO @evaluate_confidence: Dropped ratios are: [14.45, 14.94, 15.48, 16.07, 16.63, 17.1, 17.62, 18.22, 18.73, 19.28, 19.8, 20.32, 20.91, 21.5, 22.01, 22.59, 23.16, 23.72, 24.35, 25.0, 25.6, 26.24, 26.91, 27.57, 28.19, 28.86, 29.63, 30.29, 31.11, 31.82, 32.68, 33.47, 34.34, 35.1, 35.89, 36.75, 37.57, 38.41] 2020-02-01 19:25:19.586: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:25:19.586: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.12 2020-02-01 19:25:19.586: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 19:25:19.586: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 19:25:19.718: INFO @evaluate_confidence: Previous accuracy would be: 51.85 2020-02-01 19:25:19.718: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 19:25:19.720: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.8, 58.75, 59.12, 59.36, 59.38, 59.65] 2020-02-01 19:25:19.720: INFO @evaluate_confidence: Dropped ratios are: [45.65, 49.71, 53.57, 57.1, 60.37, 63.92] 2020-02-01 19:25:19.770: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:25:20.456: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:25:20.537: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:25:20.987: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:25:21.225: INFO @decay_lr : LR updated to `7.255664e-05` 2020-02-01 19:25:21.227: INFO @log_profile : T train: 121.408354 2020-02-01 19:25:21.227: INFO @log_profile : T valid: 5.460399 2020-02-01 19:25:21.227: INFO @log_profile : T read data: 2.900607 2020-02-01 19:25:21.227: INFO @log_profile : T hooks: 6.969682 2020-02-01 19:25:21.227: INFO @main_loop : Epoch 64 done 2020-02-01 19:25:21.227: INFO @main_loop : Training epoch 65 2020-02-01 19:27:40.135: INFO @log_variables: train loss nanmean: 0.803059 2020-02-01 19:27:40.136: INFO @log_variables: train age_loss mean: 5.810294 2020-02-01 19:27:40.136: INFO @log_variables: train gender_loss mean: 0.163698 2020-02-01 19:27:40.136: INFO @log_variables: train age_mae mean: 6.288191 2020-02-01 19:27:40.136: INFO @log_variables: train gender_accuracy mean: 0.932536 2020-02-01 19:27:40.136: INFO @log_variables: train gender_confidence/loss nanmean: 0.058483 2020-02-01 19:27:40.136: INFO @log_variables: train gender_confidence/accuracy mean: 0.830377 2020-02-01 19:27:40.136: INFO @log_variables: train age_confidence/loss mean: 0.067551 2020-02-01 19:27:40.136: INFO @log_variables: train age_confidence/accuracy mean: 0.606952 2020-02-01 19:27:40.136: INFO @log_variables: valid loss nanmean: 0.851253 2020-02-01 19:27:40.136: INFO @log_variables: valid age_loss mean: 5.871310 2020-02-01 19:27:40.136: INFO @log_variables: valid gender_loss mean: 0.213864 2020-02-01 19:27:40.136: INFO @log_variables: valid age_mae mean: 6.351286 2020-02-01 19:27:40.136: INFO @log_variables: valid gender_accuracy mean: 0.911673 2020-02-01 19:27:40.136: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053365 2020-02-01 19:27:40.136: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864108 2020-02-01 19:27:40.136: INFO @log_variables: valid age_confidence/loss mean: 0.069711 2020-02-01 19:27:40.136: INFO @log_variables: valid age_confidence/accuracy mean: 0.552922 2020-02-01 19:27:40.136: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:27:40.144: INFO @metrics_hook: train age_mae: 6.288 +-0.036 (110592) 2020-02-01 19:27:40.151: INFO @metrics_hook: train gender_accuracy: 0.933 +-0.002 (110592) 2020-02-01 19:27:42.848: INFO @metrics_hook: valid age_mae: 6.351 +-0.090 (17639) 2020-02-01 19:27:42.849: INFO @metrics_hook: valid gender_accuracy: 0.912 +-0.004 (17639) 2020-02-01 19:27:44.489: INFO @decay_lr : LR updated to `7.2193856e-05` 2020-02-01 19:27:44.491: INFO @log_profile : T train: 130.138303 2020-02-01 19:27:44.491: INFO @log_profile : T valid: 6.248935 2020-02-01 19:27:44.491: INFO @log_profile : T read data: 1.829472 2020-02-01 19:27:44.491: INFO @log_profile : T hooks: 4.971932 2020-02-01 19:27:44.491: INFO @main_loop : Epoch 65 done 2020-02-01 19:27:44.491: INFO @main_loop : Training epoch 66 2020-02-01 19:30:02.395: INFO @log_variables: train loss nanmean: 0.797640 2020-02-01 19:30:02.396: INFO @log_variables: train age_loss mean: 5.762000 2020-02-01 19:30:02.396: INFO @log_variables: train gender_loss mean: 0.162385 2020-02-01 19:30:02.396: INFO @log_variables: train age_mae mean: 6.240662 2020-02-01 19:30:02.396: INFO @log_variables: train gender_accuracy mean: 0.933597 2020-02-01 19:30:02.396: INFO @log_variables: train gender_confidence/loss nanmean: 0.058353 2020-02-01 19:30:02.396: INFO @log_variables: train gender_confidence/accuracy mean: 0.831452 2020-02-01 19:30:02.396: INFO @log_variables: train age_confidence/loss mean: 0.067846 2020-02-01 19:30:02.396: INFO @log_variables: train age_confidence/accuracy mean: 0.602743 2020-02-01 19:30:02.396: INFO @log_variables: valid loss nanmean: 0.842010 2020-02-01 19:30:02.396: INFO @log_variables: valid age_loss mean: 5.863356 2020-02-01 19:30:02.396: INFO @log_variables: valid gender_loss mean: 0.203030 2020-02-01 19:30:02.396: INFO @log_variables: valid age_mae mean: 6.344000 2020-02-01 19:30:02.396: INFO @log_variables: valid gender_accuracy mean: 0.914678 2020-02-01 19:30:02.396: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054615 2020-02-01 19:30:02.396: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866149 2020-02-01 19:30:02.396: INFO @log_variables: valid age_confidence/loss mean: 0.069790 2020-02-01 19:30:02.396: INFO @log_variables: valid age_confidence/accuracy mean: 0.544702 2020-02-01 19:30:02.396: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:30:02.404: INFO @metrics_hook: train age_mae: 6.241 +-0.036 (110372) 2020-02-01 19:30:02.411: INFO @metrics_hook: train gender_accuracy: 0.934 +-0.001 (110372) 2020-02-01 19:30:05.140: INFO @metrics_hook: valid age_mae: 6.344 +-0.089 (17639) 2020-02-01 19:30:05.141: INFO @metrics_hook: valid gender_accuracy: 0.915 +-0.004 (17639) 2020-02-01 19:30:06.630: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:30:06.630: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:30:06.631: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.40 +- 0.20 2020-02-01 19:30:06.631: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:30:06.762: INFO @evaluate_confidence: Previous accuracy would be: 93.36 2020-02-01 19:30:06.763: INFO @evaluate_confidence: Possible optimal thresholds are: [0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:30:06.823: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.01, 97.11, 97.19, 97.29, 97.39, 97.47, 97.56, 97.66, 97.73, 97.8, 97.88, 97.96, 98.02, 98.1, 98.15, 98.22, 98.3, 98.34, 98.4, 98.47, 98.52, 98.58, 98.63, 98.67, 98.73, 98.78, 98.83, 98.86, 98.91, 98.95, 99.0, 99.04, 99.07, 99.11, 99.15, 99.18, 99.21, 99.24, 99.27, 99.32, 99.34] 2020-02-01 19:30:06.824: INFO @evaluate_confidence: Dropped ratios are: [14.45, 15.02, 15.58, 16.14, 16.72, 17.28, 17.83, 18.4, 18.96, 19.51, 20.11, 20.66, 21.24, 21.78, 22.32, 22.88, 23.46, 23.97, 24.58, 25.16, 25.73, 26.33, 26.92, 27.46, 28.04, 28.64, 29.25, 29.89, 30.51, 31.14, 31.79, 32.45, 33.19, 33.9, 34.65, 35.36, 36.1, 36.92, 37.76, 38.55, 39.42] 2020-02-01 19:30:06.873: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:30:06.874: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.14 2020-02-01 19:30:06.874: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:30:06.874: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:30:07.012: INFO @evaluate_confidence: Previous accuracy would be: 53.54 2020-02-01 19:30:07.012: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:30:07.029: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [62.49, 63.11, 63.82, 64.4, 65.09, 65.78, 66.46, 67.32] 2020-02-01 19:30:07.029: INFO @evaluate_confidence: Dropped ratios are: [43.98, 47.0, 50.02, 52.96, 55.86, 58.65, 61.31, 63.83] 2020-02-01 19:30:07.037: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:30:07.037: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.22 2020-02-01 19:30:07.037: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.21 2020-02-01 19:30:07.037: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.24 2020-02-01 19:30:07.139: INFO @evaluate_confidence: Previous accuracy would be: 91.47 2020-02-01 19:30:07.139: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 19:30:07.148: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.96, 96.08, 96.19, 96.28, 96.43, 96.55, 96.6, 96.71, 96.82, 96.9, 97.01, 97.17, 97.24, 97.34, 97.44, 97.51, 97.6, 97.71, 97.8, 97.85, 97.92, 97.99, 98.05, 98.1, 98.16, 98.21, 98.24, 98.32, 98.39, 98.44, 98.49, 98.52, 98.58, 98.62, 98.71, 98.78, 98.85, 98.94, 99.01] 2020-02-01 19:30:07.148: INFO @evaluate_confidence: Dropped ratios are: [13.62, 14.2, 14.69, 15.15, 15.69, 16.13, 16.58, 17.17, 17.64, 18.15, 18.69, 19.22, 19.76, 20.34, 20.85, 21.33, 21.79, 22.27, 22.9, 23.36, 23.87, 24.42, 24.95, 25.59, 26.17, 26.71, 27.29, 27.97, 28.7, 29.33, 29.95, 30.55, 31.34, 32.1, 32.88, 33.63, 34.35, 35.21, 36.08] 2020-02-01 19:30:07.155: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:30:07.156: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.49 +- 0.11 2020-02-01 19:30:07.156: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.10 2020-02-01 19:30:07.156: INFO @evaluate_confidence: Average confidence of all samples 0.47 +- 0.11 2020-02-01 19:30:07.285: INFO @evaluate_confidence: Previous accuracy would be: 52.62 2020-02-01 19:30:07.285: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49] 2020-02-01 19:30:07.287: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.57, 59.79, 59.63, 59.7, 60.17] 2020-02-01 19:30:07.287: INFO @evaluate_confidence: Dropped ratios are: [47.34, 51.81, 56.03, 59.95, 63.77] 2020-02-01 19:30:07.340: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:30:08.036: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:30:08.117: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:30:08.589: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:30:08.823: INFO @decay_lr : LR updated to `7.183289e-05` 2020-02-01 19:30:08.824: INFO @log_profile : T train: 128.893342 2020-02-01 19:30:08.824: INFO @log_profile : T valid: 5.563391 2020-02-01 19:30:08.824: INFO @log_profile : T read data: 2.775794 2020-02-01 19:30:08.824: INFO @log_profile : T hooks: 7.026641 2020-02-01 19:30:08.824: INFO @main_loop : Epoch 66 done 2020-02-01 19:30:08.824: INFO @main_loop : Training epoch 67 2020-02-01 19:32:20.637: INFO @log_variables: train loss nanmean: 0.794027 2020-02-01 19:32:20.637: INFO @log_variables: train age_loss mean: 5.738282 2020-02-01 19:32:20.637: INFO @log_variables: train gender_loss mean: 0.160657 2020-02-01 19:32:20.637: INFO @log_variables: train age_mae mean: 6.215813 2020-02-01 19:32:20.637: INFO @log_variables: train gender_accuracy mean: 0.934440 2020-02-01 19:32:20.637: INFO @log_variables: train gender_confidence/loss nanmean: 0.058344 2020-02-01 19:32:20.637: INFO @log_variables: train gender_confidence/accuracy mean: 0.830682 2020-02-01 19:32:20.637: INFO @log_variables: train age_confidence/loss mean: 0.067970 2020-02-01 19:32:20.638: INFO @log_variables: train age_confidence/accuracy mean: 0.604737 2020-02-01 19:32:20.638: INFO @log_variables: valid loss nanmean: 0.849975 2020-02-01 19:32:20.638: INFO @log_variables: valid age_loss mean: 6.007263 2020-02-01 19:32:20.638: INFO @log_variables: valid gender_loss mean: 0.196157 2020-02-01 19:32:20.638: INFO @log_variables: valid age_mae mean: 6.487660 2020-02-01 19:32:20.638: INFO @log_variables: valid gender_accuracy mean: 0.919383 2020-02-01 19:32:20.638: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056177 2020-02-01 19:32:20.638: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871365 2020-02-01 19:32:20.638: INFO @log_variables: valid age_confidence/loss mean: 0.069358 2020-02-01 19:32:20.638: INFO @log_variables: valid age_confidence/accuracy mean: 0.548444 2020-02-01 19:32:20.638: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:32:20.646: INFO @metrics_hook: train age_mae: 6.216 +-0.036 (110372) 2020-02-01 19:32:20.653: INFO @metrics_hook: train gender_accuracy: 0.934 +-0.001 (110372) 2020-02-01 19:32:23.371: INFO @metrics_hook: valid age_mae: 6.488 +-0.092 (17639) 2020-02-01 19:32:23.373: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 19:32:24.991: INFO @decay_lr : LR updated to `7.1473725e-05` 2020-02-01 19:32:25.294: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 19:32:25.297: INFO @log_profile : T train: 122.870435 2020-02-01 19:32:25.297: INFO @log_profile : T valid: 5.441905 2020-02-01 19:32:25.297: INFO @log_profile : T read data: 2.820512 2020-02-01 19:32:25.297: INFO @log_profile : T hooks: 5.263380 2020-02-01 19:32:25.297: INFO @main_loop : Epoch 67 done 2020-02-01 19:32:25.297: INFO @main_loop : Training epoch 68 2020-02-01 19:34:36.302: INFO @log_variables: train loss nanmean: 0.790848 2020-02-01 19:34:36.303: INFO @log_variables: train age_loss mean: 5.717755 2020-02-01 19:34:36.303: INFO @log_variables: train gender_loss mean: 0.159275 2020-02-01 19:34:36.303: INFO @log_variables: train age_mae mean: 6.195817 2020-02-01 19:34:36.303: INFO @log_variables: train gender_accuracy mean: 0.935710 2020-02-01 19:34:36.303: INFO @log_variables: train gender_confidence/loss nanmean: 0.058305 2020-02-01 19:34:36.303: INFO @log_variables: train gender_confidence/accuracy mean: 0.830801 2020-02-01 19:34:36.303: INFO @log_variables: train age_confidence/loss mean: 0.067952 2020-02-01 19:34:36.303: INFO @log_variables: train age_confidence/accuracy mean: 0.605306 2020-02-01 19:34:36.303: INFO @log_variables: valid loss nanmean: 0.836837 2020-02-01 19:34:36.303: INFO @log_variables: valid age_loss mean: 5.872775 2020-02-01 19:34:36.303: INFO @log_variables: valid gender_loss mean: 0.197064 2020-02-01 19:34:36.303: INFO @log_variables: valid age_mae mean: 6.352964 2020-02-01 19:34:36.303: INFO @log_variables: valid gender_accuracy mean: 0.918816 2020-02-01 19:34:36.303: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053838 2020-02-01 19:34:36.303: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864618 2020-02-01 19:34:36.303: INFO @log_variables: valid age_confidence/loss mean: 0.069961 2020-02-01 19:34:36.304: INFO @log_variables: valid age_confidence/accuracy mean: 0.557231 2020-02-01 19:34:36.304: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:34:36.311: INFO @metrics_hook: train age_mae: 6.196 +-0.035 (110592) 2020-02-01 19:34:36.318: INFO @metrics_hook: train gender_accuracy: 0.936 +-0.001 (110592) 2020-02-01 19:34:39.095: INFO @metrics_hook: valid age_mae: 6.353 +-0.090 (17639) 2020-02-01 19:34:39.096: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 19:34:40.570: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:34:40.570: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:34:40.570: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.20 2020-02-01 19:34:40.570: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:34:40.699: INFO @evaluate_confidence: Previous accuracy would be: 93.57 2020-02-01 19:34:40.700: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:34:40.760: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.02, 97.12, 97.25, 97.33, 97.42, 97.5, 97.59, 97.67, 97.76, 97.85, 97.91, 97.98, 98.04, 98.1, 98.16, 98.22, 98.28, 98.35, 98.43, 98.48, 98.54, 98.59, 98.64, 98.67, 98.72, 98.77, 98.81, 98.85, 98.88, 98.92, 98.98, 99.02, 99.06, 99.11, 99.15, 99.19, 99.23, 99.26, 99.29, 99.31, 99.35, 99.38] 2020-02-01 19:34:40.760: INFO @evaluate_confidence: Dropped ratios are: [13.99, 14.59, 15.2, 15.74, 16.29, 16.77, 17.37, 17.95, 18.48, 19.04, 19.59, 20.13, 20.7, 21.26, 21.84, 22.36, 22.93, 23.46, 24.06, 24.62, 25.2, 25.79, 26.35, 26.93, 27.55, 28.14, 28.75, 29.35, 29.96, 30.58, 31.23, 31.91, 32.56, 33.23, 33.92, 34.65, 35.37, 36.08, 36.84, 37.66, 38.52, 39.35] 2020-02-01 19:34:40.809: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:34:40.809: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.14 2020-02-01 19:34:40.810: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:34:40.810: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:34:40.944: INFO @evaluate_confidence: Previous accuracy would be: 54.01 2020-02-01 19:34:40.944: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:34:40.960: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.14, 63.75, 64.42, 65.08, 65.8, 66.56, 67.28, 68.08] 2020-02-01 19:34:40.961: INFO @evaluate_confidence: Dropped ratios are: [43.56, 46.59, 49.61, 52.59, 55.5, 58.26, 60.93, 63.45] 2020-02-01 19:34:40.968: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:34:40.968: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.22 2020-02-01 19:34:40.968: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.21 2020-02-01 19:34:40.968: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.24 2020-02-01 19:34:41.067: INFO @evaluate_confidence: Previous accuracy would be: 91.88 2020-02-01 19:34:41.068: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 19:34:41.075: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.03, 96.16, 96.29, 96.41, 96.53, 96.67, 96.83, 96.96, 97.06, 97.17, 97.25, 97.37, 97.48, 97.56, 97.65, 97.69, 97.74, 97.84, 97.92, 97.99, 98.03, 98.11, 98.18, 98.25, 98.35, 98.41, 98.47, 98.52, 98.54, 98.56, 98.6, 98.65, 98.67, 98.7, 98.73, 98.83, 98.89] 2020-02-01 19:34:41.076: INFO @evaluate_confidence: Dropped ratios are: [13.8, 14.29, 14.83, 15.3, 15.82, 16.3, 16.85, 17.5, 17.97, 18.48, 18.95, 19.52, 20.06, 20.62, 21.23, 21.72, 22.34, 23.0, 23.58, 24.13, 24.63, 25.17, 25.68, 26.3, 26.95, 27.63, 28.25, 28.9, 29.58, 30.27, 31.03, 31.75, 32.41, 33.1, 33.98, 34.89, 35.72] 2020-02-01 19:34:41.083: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:34:41.083: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 19:34:41.083: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 19:34:41.084: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 19:34:41.208: INFO @evaluate_confidence: Previous accuracy would be: 52.83 2020-02-01 19:34:41.208: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 19:34:41.210: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.12, 59.19, 59.24, 59.78] 2020-02-01 19:34:41.210: INFO @evaluate_confidence: Dropped ratios are: [44.88, 48.65, 52.57, 56.29] 2020-02-01 19:34:41.260: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:34:41.957: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:34:42.039: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:34:46.544: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:34:46.783: INFO @decay_lr : LR updated to `7.111636e-05` 2020-02-01 19:34:46.784: INFO @log_profile : T train: 122.861590 2020-02-01 19:34:46.784: INFO @log_profile : T valid: 5.468316 2020-02-01 19:34:46.784: INFO @log_profile : T read data: 1.977996 2020-02-01 19:34:46.784: INFO @log_profile : T hooks: 11.100116 2020-02-01 19:34:46.784: INFO @main_loop : Epoch 68 done 2020-02-01 19:34:46.784: INFO @main_loop : Training epoch 69 2020-02-01 19:37:07.053: INFO @log_variables: train loss nanmean: 0.787034 2020-02-01 19:37:07.053: INFO @log_variables: train age_loss mean: 5.696568 2020-02-01 19:37:07.053: INFO @log_variables: train gender_loss mean: 0.157680 2020-02-01 19:37:07.053: INFO @log_variables: train age_mae mean: 6.174656 2020-02-01 19:37:07.053: INFO @log_variables: train gender_accuracy mean: 0.936270 2020-02-01 19:37:07.053: INFO @log_variables: train gender_confidence/loss nanmean: 0.057815 2020-02-01 19:37:07.053: INFO @log_variables: train gender_confidence/accuracy mean: 0.833644 2020-02-01 19:37:07.053: INFO @log_variables: train age_confidence/loss mean: 0.068004 2020-02-01 19:37:07.053: INFO @log_variables: train age_confidence/accuracy mean: 0.604247 2020-02-01 19:37:07.054: INFO @log_variables: valid loss nanmean: 0.862279 2020-02-01 19:37:07.054: INFO @log_variables: valid age_loss mean: 5.946331 2020-02-01 19:37:07.054: INFO @log_variables: valid gender_loss mean: 0.216373 2020-02-01 19:37:07.054: INFO @log_variables: valid age_mae mean: 6.427387 2020-02-01 19:37:07.054: INFO @log_variables: valid gender_accuracy mean: 0.915415 2020-02-01 19:37:07.054: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055275 2020-02-01 19:37:07.054: INFO @log_variables: valid gender_confidence/accuracy mean: 0.863995 2020-02-01 19:37:07.054: INFO @log_variables: valid age_confidence/loss mean: 0.069726 2020-02-01 19:37:07.054: INFO @log_variables: valid age_confidence/accuracy mean: 0.559782 2020-02-01 19:37:07.054: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:37:07.061: INFO @metrics_hook: train age_mae: 6.175 +-0.035 (110372) 2020-02-01 19:37:07.068: INFO @metrics_hook: train gender_accuracy: 0.936 +-0.001 (110372) 2020-02-01 19:37:09.868: INFO @metrics_hook: valid age_mae: 6.427 +-0.092 (17639) 2020-02-01 19:37:09.870: INFO @metrics_hook: valid gender_accuracy: 0.915 +-0.004 (17639) 2020-02-01 19:37:11.536: INFO @decay_lr : LR updated to `7.0760776e-05` 2020-02-01 19:37:11.537: INFO @log_profile : T train: 130.557145 2020-02-01 19:37:11.537: INFO @log_profile : T valid: 6.262023 2020-02-01 19:37:11.537: INFO @log_profile : T read data: 2.784522 2020-02-01 19:37:11.537: INFO @log_profile : T hooks: 5.073443 2020-02-01 19:37:11.537: INFO @main_loop : Epoch 69 done 2020-02-01 19:37:11.537: INFO @main_loop : Training epoch 70 2020-02-01 19:39:32.184: INFO @log_variables: train loss nanmean: 0.789023 2020-02-01 19:39:32.184: INFO @log_variables: train age_loss mean: 5.703395 2020-02-01 19:39:32.185: INFO @log_variables: train gender_loss mean: 0.159527 2020-02-01 19:39:32.185: INFO @log_variables: train age_mae mean: 6.181187 2020-02-01 19:39:32.185: INFO @log_variables: train gender_accuracy mean: 0.935319 2020-02-01 19:39:32.185: INFO @log_variables: train gender_confidence/loss nanmean: 0.057466 2020-02-01 19:39:32.185: INFO @log_variables: train gender_confidence/accuracy mean: 0.832004 2020-02-01 19:39:32.185: INFO @log_variables: train age_confidence/loss mean: 0.068042 2020-02-01 19:39:32.185: INFO @log_variables: train age_confidence/accuracy mean: 0.604537 2020-02-01 19:39:32.185: INFO @log_variables: valid loss nanmean: 0.843110 2020-02-01 19:39:32.185: INFO @log_variables: valid age_loss mean: 5.854396 2020-02-01 19:39:32.185: INFO @log_variables: valid gender_loss mean: 0.205234 2020-02-01 19:39:32.185: INFO @log_variables: valid age_mae mean: 6.334221 2020-02-01 19:39:32.185: INFO @log_variables: valid gender_accuracy mean: 0.917626 2020-02-01 19:39:32.185: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054794 2020-02-01 19:39:32.185: INFO @log_variables: valid gender_confidence/accuracy mean: 0.863995 2020-02-01 19:39:32.185: INFO @log_variables: valid age_confidence/loss mean: 0.069522 2020-02-01 19:39:32.185: INFO @log_variables: valid age_confidence/accuracy mean: 0.560859 2020-02-01 19:39:32.185: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:39:32.193: INFO @metrics_hook: train age_mae: 6.181 +-0.035 (110372) 2020-02-01 19:39:32.200: INFO @metrics_hook: train gender_accuracy: 0.935 +-0.001 (110372) 2020-02-01 19:39:34.961: INFO @metrics_hook: valid age_mae: 6.334 +-0.090 (17639) 2020-02-01 19:39:34.963: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 19:39:36.446: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:39:36.446: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:39:36.446: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.20 2020-02-01 19:39:36.447: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:39:36.580: INFO @evaluate_confidence: Previous accuracy would be: 93.53 2020-02-01 19:39:36.580: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:39:36.643: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.06, 97.16, 97.26, 97.35, 97.43, 97.54, 97.62, 97.71, 97.79, 97.87, 97.95, 98.02, 98.09, 98.15, 98.21, 98.28, 98.34, 98.41, 98.47, 98.51, 98.57, 98.63, 98.68, 98.72, 98.77, 98.81, 98.85, 98.89, 98.94, 98.98, 99.03, 99.07, 99.09, 99.13, 99.16, 99.2, 99.23, 99.27, 99.29, 99.33, 99.35, 99.37] 2020-02-01 19:39:36.643: INFO @evaluate_confidence: Dropped ratios are: [14.06, 14.62, 15.18, 15.73, 16.27, 16.81, 17.38, 17.92, 18.45, 18.97, 19.54, 20.1, 20.65, 21.23, 21.79, 22.33, 22.9, 23.44, 24.0, 24.6, 25.15, 25.72, 26.27, 26.83, 27.42, 27.97, 28.59, 29.21, 29.78, 30.39, 30.96, 31.63, 32.27, 32.92, 33.56, 34.26, 34.99, 35.72, 36.49, 37.27, 38.03, 38.84] 2020-02-01 19:39:36.694: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:39:36.694: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.14 2020-02-01 19:39:36.694: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:39:36.694: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:39:36.833: INFO @evaluate_confidence: Previous accuracy would be: 54.04 2020-02-01 19:39:36.834: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:39:36.851: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.15, 63.85, 64.46, 65.22, 65.9, 66.7, 67.48, 68.32] 2020-02-01 19:39:36.851: INFO @evaluate_confidence: Dropped ratios are: [43.77, 46.87, 49.9, 52.87, 55.67, 58.44, 61.16, 63.75] 2020-02-01 19:39:36.859: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:39:36.859: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.21 2020-02-01 19:39:36.860: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.21 2020-02-01 19:39:36.860: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.24 2020-02-01 19:39:36.966: INFO @evaluate_confidence: Previous accuracy would be: 91.76 2020-02-01 19:39:36.967: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 19:39:36.975: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.84, 95.89, 96.01, 96.1, 96.27, 96.35, 96.47, 96.58, 96.73, 96.83, 96.92, 97.03, 97.12, 97.26, 97.35, 97.43, 97.48, 97.58, 97.69, 97.78, 97.85, 97.88, 97.93, 97.97, 98.05, 98.11, 98.22, 98.31, 98.37, 98.45, 98.53, 98.55, 98.62, 98.71, 98.77, 98.88, 98.96] 2020-02-01 19:39:36.975: INFO @evaluate_confidence: Dropped ratios are: [13.83, 14.29, 14.71, 15.23, 15.75, 16.2, 16.62, 17.09, 17.61, 18.1, 18.54, 19.17, 19.63, 20.15, 20.64, 21.12, 21.66, 22.29, 22.79, 23.27, 23.78, 24.27, 24.88, 25.44, 26.06, 26.65, 27.33, 28.0, 28.69, 29.47, 30.18, 30.9, 31.69, 32.48, 33.31, 34.23, 35.17] 2020-02-01 19:39:36.983: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:39:36.983: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 19:39:36.983: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.09 2020-02-01 19:39:36.983: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 19:39:37.112: INFO @evaluate_confidence: Previous accuracy would be: 52.45 2020-02-01 19:39:37.112: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 19:39:37.113: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.21, 59.47, 59.88, 60.04] 2020-02-01 19:39:37.113: INFO @evaluate_confidence: Dropped ratios are: [43.53, 48.36, 53.14, 57.52] 2020-02-01 19:39:37.168: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:39:37.874: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:39:37.958: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:39:38.419: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:39:38.498: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:39:39.178: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:39:39.261: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 19:39:39.263: INFO @evaluate_gender-age_model: groups 0 4.282735 1 4.771572 2 5.643043 3 5.936015 4 6.844545 5 6.842388 6 6.930076 7 7.976628 Name: errors, dtype: float64 2020-02-01 19:39:39.264: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:39:39.712: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:39:39.774: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 19:39:39.776: INFO @evaluate_gender-age_model: groups 0 6.573656 1 5.605777 2 5.563028 3 5.562642 4 7.401115 5 5.581087 6 7.740327 7 12.899268 Name: errors, dtype: float64 2020-02-01 19:39:39.948: INFO @decay_lr : LR updated to `7.0406975e-05` 2020-02-01 19:39:39.949: INFO @log_profile : T train: 129.846580 2020-02-01 19:39:39.950: INFO @log_profile : T valid: 5.411743 2020-02-01 19:39:39.950: INFO @log_profile : T read data: 2.817971 2020-02-01 19:39:39.950: INFO @log_profile : T hooks: 10.260316 2020-02-01 19:39:39.950: INFO @main_loop : Epoch 70 done 2020-02-01 19:39:39.950: INFO @main_loop : Training epoch 71 2020-02-01 19:41:50.068: INFO @log_variables: train loss nanmean: 0.785751 2020-02-01 19:41:50.068: INFO @log_variables: train age_loss mean: 5.677046 2020-02-01 19:41:50.069: INFO @log_variables: train gender_loss mean: 0.157974 2020-02-01 19:41:50.069: INFO @log_variables: train age_mae mean: 6.154788 2020-02-01 19:41:50.069: INFO @log_variables: train gender_accuracy mean: 0.936062 2020-02-01 19:41:50.069: INFO @log_variables: train gender_confidence/loss nanmean: 0.057893 2020-02-01 19:41:50.069: INFO @log_variables: train gender_confidence/accuracy mean: 0.833324 2020-02-01 19:41:50.069: INFO @log_variables: train age_confidence/loss mean: 0.068150 2020-02-01 19:41:50.069: INFO @log_variables: train age_confidence/accuracy mean: 0.605903 2020-02-01 19:41:50.069: INFO @log_variables: valid loss nanmean: 0.840224 2020-02-01 19:41:50.069: INFO @log_variables: valid age_loss mean: 5.875483 2020-02-01 19:41:50.069: INFO @log_variables: valid gender_loss mean: 0.201814 2020-02-01 19:41:50.069: INFO @log_variables: valid age_mae mean: 6.356172 2020-02-01 19:41:50.069: INFO @log_variables: valid gender_accuracy mean: 0.916889 2020-02-01 19:41:50.069: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053280 2020-02-01 19:41:50.069: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862804 2020-02-01 19:41:50.069: INFO @log_variables: valid age_confidence/loss mean: 0.069343 2020-02-01 19:41:50.069: INFO @log_variables: valid age_confidence/accuracy mean: 0.572368 2020-02-01 19:41:50.069: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:41:50.076: INFO @metrics_hook: train age_mae: 6.155 +-0.035 (110592) 2020-02-01 19:41:50.083: INFO @metrics_hook: train gender_accuracy: 0.936 +-0.001 (110592) 2020-02-01 19:41:52.837: INFO @metrics_hook: valid age_mae: 6.356 +-0.090 (17639) 2020-02-01 19:41:52.839: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 19:41:54.488: INFO @decay_lr : LR updated to `7.005494e-05` 2020-02-01 19:41:54.489: INFO @log_profile : T train: 121.990912 2020-02-01 19:41:54.489: INFO @log_profile : T valid: 5.452413 2020-02-01 19:41:54.489: INFO @log_profile : T read data: 1.965051 2020-02-01 19:41:54.490: INFO @log_profile : T hooks: 5.051890 2020-02-01 19:41:54.490: INFO @main_loop : Epoch 71 done 2020-02-01 19:41:54.490: INFO @main_loop : Training epoch 72 2020-02-01 19:44:14.324: INFO @log_variables: train loss nanmean: 0.784452 2020-02-01 19:44:14.325: INFO @log_variables: train age_loss mean: 5.668228 2020-02-01 19:44:14.325: INFO @log_variables: train gender_loss mean: 0.157391 2020-02-01 19:44:14.325: INFO @log_variables: train age_mae mean: 6.146072 2020-02-01 19:44:14.325: INFO @log_variables: train gender_accuracy mean: 0.936741 2020-02-01 19:44:14.325: INFO @log_variables: train gender_confidence/loss nanmean: 0.057875 2020-02-01 19:44:14.325: INFO @log_variables: train gender_confidence/accuracy mean: 0.833273 2020-02-01 19:44:14.325: INFO @log_variables: train age_confidence/loss mean: 0.068200 2020-02-01 19:44:14.325: INFO @log_variables: train age_confidence/accuracy mean: 0.604592 2020-02-01 19:44:14.325: INFO @log_variables: valid loss nanmean: 0.846151 2020-02-01 19:44:14.325: INFO @log_variables: valid age_loss mean: 5.891201 2020-02-01 19:44:14.325: INFO @log_variables: valid gender_loss mean: 0.203935 2020-02-01 19:44:14.325: INFO @log_variables: valid age_mae mean: 6.372416 2020-02-01 19:44:14.325: INFO @log_variables: valid gender_accuracy mean: 0.916832 2020-02-01 19:44:14.325: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055630 2020-02-01 19:44:14.325: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868133 2020-02-01 19:44:14.325: INFO @log_variables: valid age_confidence/loss mean: 0.069562 2020-02-01 19:44:14.325: INFO @log_variables: valid age_confidence/accuracy mean: 0.558138 2020-02-01 19:44:14.325: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:44:14.332: INFO @metrics_hook: train age_mae: 6.146 +-0.035 (110372) 2020-02-01 19:44:14.340: INFO @metrics_hook: train gender_accuracy: 0.937 +-0.001 (110372) 2020-02-01 19:44:17.082: INFO @metrics_hook: valid age_mae: 6.372 +-0.089 (17639) 2020-02-01 19:44:17.083: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 19:44:18.558: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:44:18.559: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:44:18.559: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.39 +- 0.20 2020-02-01 19:44:18.559: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:44:18.689: INFO @evaluate_confidence: Previous accuracy would be: 93.67 2020-02-01 19:44:18.689: INFO @evaluate_confidence: Possible optimal thresholds are: [0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:44:18.751: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.1, 97.2, 97.31, 97.42, 97.5, 97.58, 97.66, 97.74, 97.83, 97.88, 97.95, 98.01, 98.08, 98.15, 98.21, 98.27, 98.33, 98.4, 98.44, 98.49, 98.56, 98.63, 98.68, 98.73, 98.79, 98.83, 98.89, 98.94, 98.98, 99.01, 99.06, 99.07, 99.11, 99.13, 99.17, 99.2, 99.21, 99.26, 99.29, 99.34, 99.37, 99.39] 2020-02-01 19:44:18.751: INFO @evaluate_confidence: Dropped ratios are: [13.82, 14.38, 14.96, 15.51, 16.02, 16.58, 17.09, 17.63, 18.17, 18.68, 19.28, 19.81, 20.37, 20.92, 21.43, 21.97, 22.5, 23.05, 23.6, 24.14, 24.71, 25.28, 25.85, 26.48, 27.08, 27.65, 28.23, 28.79, 29.42, 30.02, 30.67, 31.29, 31.96, 32.62, 33.32, 34.0, 34.68, 35.42, 36.17, 36.95, 37.71, 38.55] 2020-02-01 19:44:18.802: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:44:18.802: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.14 2020-02-01 19:44:18.802: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:44:18.803: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:44:18.943: INFO @evaluate_confidence: Previous accuracy would be: 54.14 2020-02-01 19:44:18.943: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 19:44:18.960: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.1, 63.69, 64.38, 65.08, 65.87, 66.54, 67.36, 68.25] 2020-02-01 19:44:18.960: INFO @evaluate_confidence: Dropped ratios are: [43.16, 46.21, 49.23, 52.37, 55.32, 58.05, 60.77, 63.34] 2020-02-01 19:44:18.968: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:44:18.968: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.21 2020-02-01 19:44:18.968: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.22 2020-02-01 19:44:18.968: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.24 2020-02-01 19:44:19.073: INFO @evaluate_confidence: Previous accuracy would be: 91.68 2020-02-01 19:44:19.074: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 19:44:19.081: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.84, 95.92, 96.1, 96.23, 96.32, 96.38, 96.48, 96.58, 96.67, 96.79, 96.89, 97.08, 97.19, 97.26, 97.34, 97.45, 97.53, 97.59, 97.67, 97.71, 97.81, 97.92, 97.94, 98.02, 98.07, 98.14, 98.19, 98.3, 98.33, 98.39, 98.5, 98.56, 98.59, 98.67, 98.72, 98.82, 98.87] 2020-02-01 19:44:19.082: INFO @evaluate_confidence: Dropped ratios are: [13.74, 14.18, 14.63, 15.1, 15.53, 16.01, 16.48, 16.97, 17.43, 17.94, 18.37, 18.94, 19.41, 19.9, 20.34, 20.93, 21.45, 21.95, 22.57, 23.15, 23.75, 24.4, 24.96, 25.6, 26.15, 26.78, 27.49, 28.13, 28.85, 29.52, 30.29, 31.09, 31.92, 32.73, 33.59, 34.45, 35.3] 2020-02-01 19:44:19.089: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:44:19.089: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 19:44:19.089: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 19:44:19.090: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.10 2020-02-01 19:44:19.220: INFO @evaluate_confidence: Previous accuracy would be: 52.37 2020-02-01 19:44:19.220: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 19:44:19.221: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.62, 59.82, 60.32, 60.76] 2020-02-01 19:44:19.221: INFO @evaluate_confidence: Dropped ratios are: [45.05, 49.11, 53.37, 57.61] 2020-02-01 19:44:19.277: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:44:20.015: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:44:20.100: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:44:20.571: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:44:20.814: INFO @decay_lr : LR updated to `6.970467e-05` 2020-02-01 19:44:20.816: INFO @log_profile : T train: 129.628166 2020-02-01 19:44:20.816: INFO @log_profile : T valid: 6.774333 2020-02-01 19:44:20.816: INFO @log_profile : T read data: 2.759185 2020-02-01 19:44:20.816: INFO @log_profile : T hooks: 7.088492 2020-02-01 19:44:20.816: INFO @main_loop : Epoch 72 done 2020-02-01 19:44:20.816: INFO @main_loop : Training epoch 73 2020-02-01 19:46:39.799: INFO @log_variables: train loss nanmean: 0.779736 2020-02-01 19:46:39.799: INFO @log_variables: train age_loss mean: 5.658925 2020-02-01 19:46:39.799: INFO @log_variables: train gender_loss mean: 0.153900 2020-02-01 19:46:39.799: INFO @log_variables: train age_mae mean: 6.136550 2020-02-01 19:46:39.799: INFO @log_variables: train gender_accuracy mean: 0.937738 2020-02-01 19:46:39.799: INFO @log_variables: train gender_confidence/loss nanmean: 0.057160 2020-02-01 19:46:39.799: INFO @log_variables: train gender_confidence/accuracy mean: 0.836951 2020-02-01 19:46:39.799: INFO @log_variables: train age_confidence/loss mean: 0.068219 2020-02-01 19:46:39.799: INFO @log_variables: train age_confidence/accuracy mean: 0.605969 2020-02-01 19:46:39.799: INFO @log_variables: valid loss nanmean: 0.839587 2020-02-01 19:46:39.799: INFO @log_variables: valid age_loss mean: 5.916780 2020-02-01 19:46:39.799: INFO @log_variables: valid gender_loss mean: 0.196123 2020-02-01 19:46:39.800: INFO @log_variables: valid age_mae mean: 6.397953 2020-02-01 19:46:39.800: INFO @log_variables: valid gender_accuracy mean: 0.922501 2020-02-01 19:46:39.800: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053846 2020-02-01 19:46:39.800: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876694 2020-02-01 19:46:39.800: INFO @log_variables: valid age_confidence/loss mean: 0.069558 2020-02-01 19:46:39.800: INFO @log_variables: valid age_confidence/accuracy mean: 0.555814 2020-02-01 19:46:39.800: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:46:39.807: INFO @metrics_hook: train age_mae: 6.137 +-0.035 (110372) 2020-02-01 19:46:39.814: INFO @metrics_hook: train gender_accuracy: 0.938 +-0.001 (110372) 2020-02-01 19:46:42.555: INFO @metrics_hook: valid age_mae: 6.398 +-0.090 (17639) 2020-02-01 19:46:42.556: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 19:46:44.206: INFO @decay_lr : LR updated to `6.9356145e-05` 2020-02-01 19:46:44.517: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 19:46:44.520: INFO @log_profile : T train: 129.762917 2020-02-01 19:46:44.520: INFO @log_profile : T valid: 6.730243 2020-02-01 19:46:44.520: INFO @log_profile : T read data: 1.806345 2020-02-01 19:46:44.520: INFO @log_profile : T hooks: 5.329580 2020-02-01 19:46:44.520: INFO @main_loop : Epoch 73 done 2020-02-01 19:46:44.520: INFO @main_loop : Training epoch 74 2020-02-01 19:49:02.425: INFO @log_variables: train loss nanmean: 0.781871 2020-02-01 19:49:02.425: INFO @log_variables: train age_loss mean: 5.642569 2020-02-01 19:49:02.425: INFO @log_variables: train gender_loss mean: 0.157729 2020-02-01 19:49:02.425: INFO @log_variables: train age_mae mean: 6.120032 2020-02-01 19:49:02.425: INFO @log_variables: train gender_accuracy mean: 0.935366 2020-02-01 19:49:02.425: INFO @log_variables: train gender_confidence/loss nanmean: 0.057322 2020-02-01 19:49:02.425: INFO @log_variables: train gender_confidence/accuracy mean: 0.833523 2020-02-01 19:49:02.425: INFO @log_variables: train age_confidence/loss mean: 0.068198 2020-02-01 19:49:02.425: INFO @log_variables: train age_confidence/accuracy mean: 0.608200 2020-02-01 19:49:02.425: INFO @log_variables: valid loss nanmean: 0.853840 2020-02-01 19:49:02.425: INFO @log_variables: valid age_loss mean: 5.840625 2020-02-01 19:49:02.425: INFO @log_variables: valid gender_loss mean: 0.218652 2020-02-01 19:49:02.425: INFO @log_variables: valid age_mae mean: 6.320879 2020-02-01 19:49:02.426: INFO @log_variables: valid gender_accuracy mean: 0.909575 2020-02-01 19:49:02.426: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053977 2020-02-01 19:49:02.426: INFO @log_variables: valid gender_confidence/accuracy mean: 0.855547 2020-02-01 19:49:02.426: INFO @log_variables: valid age_confidence/loss mean: 0.070122 2020-02-01 19:49:02.426: INFO @log_variables: valid age_confidence/accuracy mean: 0.551845 2020-02-01 19:49:02.426: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:49:02.433: INFO @metrics_hook: train age_mae: 6.120 +-0.035 (110592) 2020-02-01 19:49:02.440: INFO @metrics_hook: train gender_accuracy: 0.935 +-0.001 (110592) 2020-02-01 19:49:05.137: INFO @metrics_hook: valid age_mae: 6.321 +-0.089 (17639) 2020-02-01 19:49:05.139: INFO @metrics_hook: valid gender_accuracy: 0.910 +-0.004 (17639) 2020-02-01 19:49:06.592: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:49:06.592: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:49:06.592: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 19:49:06.592: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:49:06.719: INFO @evaluate_confidence: Previous accuracy would be: 93.54 2020-02-01 19:49:06.720: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:49:06.782: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.03, 97.14, 97.23, 97.34, 97.42, 97.5, 97.58, 97.65, 97.75, 97.83, 97.9, 97.97, 98.03, 98.1, 98.18, 98.24, 98.31, 98.37, 98.43, 98.49, 98.54, 98.61, 98.65, 98.69, 98.74, 98.79, 98.85, 98.89, 98.93, 98.96, 99.0, 99.04, 99.07, 99.1, 99.14, 99.17, 99.2, 99.23, 99.27, 99.3, 99.32, 99.35, 99.39] 2020-02-01 19:49:06.782: INFO @evaluate_confidence: Dropped ratios are: [13.49, 14.08, 14.62, 15.17, 15.71, 16.24, 16.78, 17.3, 17.85, 18.4, 18.9, 19.42, 19.95, 20.51, 21.04, 21.59, 22.13, 22.67, 23.22, 23.78, 24.32, 24.91, 25.45, 25.99, 26.55, 27.08, 27.7, 28.25, 28.84, 29.43, 30.06, 30.66, 31.29, 31.9, 32.53, 33.18, 33.92, 34.63, 35.4, 36.16, 36.91, 37.7, 38.52] 2020-02-01 19:49:06.830: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:49:06.830: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 19:49:06.831: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:49:06.831: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:49:06.966: INFO @evaluate_confidence: Previous accuracy would be: 54.46 2020-02-01 19:49:06.966: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 19:49:06.985: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.58, 64.19, 64.87, 65.59, 66.27, 66.95, 67.79, 68.47, 69.19] 2020-02-01 19:49:06.986: INFO @evaluate_confidence: Dropped ratios are: [42.58, 45.63, 48.66, 51.5, 54.38, 57.19, 59.88, 62.57, 65.03] 2020-02-01 19:49:06.993: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:49:06.993: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.22 2020-02-01 19:49:06.994: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.21 2020-02-01 19:49:06.994: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.25 2020-02-01 19:49:07.095: INFO @evaluate_confidence: Previous accuracy would be: 90.96 2020-02-01 19:49:07.095: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 19:49:07.104: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.49, 95.68, 95.77, 95.86, 96.04, 96.17, 96.31, 96.39, 96.52, 96.63, 96.72, 96.89, 96.94, 97.03, 97.12, 97.25, 97.33, 97.41, 97.49, 97.55, 97.62, 97.68, 97.84, 97.88, 97.99, 98.03, 98.07, 98.15, 98.19, 98.27, 98.39, 98.47, 98.49, 98.57, 98.61, 98.67, 98.76, 98.82] 2020-02-01 19:49:07.104: INFO @evaluate_confidence: Dropped ratios are: [14.33, 14.81, 15.34, 15.87, 16.4, 16.88, 17.4, 17.8, 18.33, 18.79, 19.36, 19.89, 20.51, 21.06, 21.56, 22.12, 22.61, 23.27, 23.84, 24.46, 25.06, 25.59, 26.28, 26.83, 27.5, 28.1, 28.71, 29.33, 30.04, 30.72, 31.48, 32.25, 33.07, 33.77, 34.55, 35.43, 36.36, 37.35] 2020-02-01 19:49:07.112: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:49:07.112: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 19:49:07.112: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 19:49:07.112: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.11 2020-02-01 19:49:07.238: INFO @evaluate_confidence: Previous accuracy would be: 53.01 2020-02-01 19:49:07.238: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 19:49:07.239: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.93, 59.18, 59.47, 59.71] 2020-02-01 19:49:07.239: INFO @evaluate_confidence: Dropped ratios are: [44.93, 48.59, 52.3, 55.8] 2020-02-01 19:49:07.292: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:49:07.990: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:49:08.075: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:49:08.533: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:49:08.777: INFO @decay_lr : LR updated to `6.9009366e-05` 2020-02-01 19:49:08.778: INFO @log_profile : T train: 128.775697 2020-02-01 19:49:08.778: INFO @log_profile : T valid: 5.699586 2020-02-01 19:49:08.778: INFO @log_profile : T read data: 2.732295 2020-02-01 19:49:08.778: INFO @log_profile : T hooks: 6.974583 2020-02-01 19:49:08.779: INFO @main_loop : Epoch 74 done 2020-02-01 19:49:08.779: INFO @main_loop : Training epoch 75 2020-02-01 19:51:20.573: INFO @log_variables: train loss nanmean: 0.776436 2020-02-01 19:51:20.573: INFO @log_variables: train age_loss mean: 5.629188 2020-02-01 19:51:20.573: INFO @log_variables: train gender_loss mean: 0.152996 2020-02-01 19:51:20.573: INFO @log_variables: train age_mae mean: 6.106796 2020-02-01 19:51:20.573: INFO @log_variables: train gender_accuracy mean: 0.938001 2020-02-01 19:51:20.573: INFO @log_variables: train gender_confidence/loss nanmean: 0.057279 2020-02-01 19:51:20.573: INFO @log_variables: train gender_confidence/accuracy mean: 0.834958 2020-02-01 19:51:20.573: INFO @log_variables: train age_confidence/loss mean: 0.068326 2020-02-01 19:51:20.574: INFO @log_variables: train age_confidence/accuracy mean: 0.607636 2020-02-01 19:51:20.574: INFO @log_variables: valid loss nanmean: 0.837475 2020-02-01 19:51:20.574: INFO @log_variables: valid age_loss mean: 5.840611 2020-02-01 19:51:20.574: INFO @log_variables: valid gender_loss mean: 0.200540 2020-02-01 19:51:20.574: INFO @log_variables: valid age_mae mean: 6.321509 2020-02-01 19:51:20.574: INFO @log_variables: valid gender_accuracy mean: 0.917682 2020-02-01 19:51:20.574: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054457 2020-02-01 19:51:20.574: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872895 2020-02-01 19:51:20.574: INFO @log_variables: valid age_confidence/loss mean: 0.069744 2020-02-01 19:51:20.574: INFO @log_variables: valid age_confidence/accuracy mean: 0.568400 2020-02-01 19:51:20.574: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:51:20.581: INFO @metrics_hook: train age_mae: 6.107 +-0.035 (110372) 2020-02-01 19:51:20.588: INFO @metrics_hook: train gender_accuracy: 0.938 +-0.001 (110372) 2020-02-01 19:51:23.304: INFO @metrics_hook: valid age_mae: 6.322 +-0.089 (17639) 2020-02-01 19:51:23.305: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 19:51:24.925: INFO @decay_lr : LR updated to `6.866432e-05` 2020-02-01 19:51:24.926: INFO @log_profile : T train: 122.804441 2020-02-01 19:51:24.927: INFO @log_profile : T valid: 5.475056 2020-02-01 19:51:24.927: INFO @log_profile : T read data: 2.821113 2020-02-01 19:51:24.927: INFO @log_profile : T hooks: 4.969952 2020-02-01 19:51:24.927: INFO @main_loop : Epoch 75 done 2020-02-01 19:51:24.927: INFO @main_loop : Training epoch 76 2020-02-01 19:53:35.471: INFO @log_variables: train loss nanmean: 0.772901 2020-02-01 19:53:35.471: INFO @log_variables: train age_loss mean: 5.595585 2020-02-01 19:53:35.471: INFO @log_variables: train gender_loss mean: 0.152440 2020-02-01 19:53:35.471: INFO @log_variables: train age_mae mean: 6.072961 2020-02-01 19:53:35.471: INFO @log_variables: train gender_accuracy mean: 0.938712 2020-02-01 19:53:35.471: INFO @log_variables: train gender_confidence/loss nanmean: 0.057177 2020-02-01 19:53:35.471: INFO @log_variables: train gender_confidence/accuracy mean: 0.835223 2020-02-01 19:53:35.471: INFO @log_variables: train age_confidence/loss mean: 0.068453 2020-02-01 19:53:35.471: INFO @log_variables: train age_confidence/accuracy mean: 0.605912 2020-02-01 19:53:35.471: INFO @log_variables: valid loss nanmean: 0.838084 2020-02-01 19:53:35.471: INFO @log_variables: valid age_loss mean: 5.883421 2020-02-01 19:53:35.471: INFO @log_variables: valid gender_loss mean: 0.198403 2020-02-01 19:53:35.471: INFO @log_variables: valid age_mae mean: 6.363863 2020-02-01 19:53:35.471: INFO @log_variables: valid gender_accuracy mean: 0.916378 2020-02-01 19:53:35.471: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053058 2020-02-01 19:53:35.472: INFO @log_variables: valid gender_confidence/accuracy mean: 0.856171 2020-02-01 19:53:35.472: INFO @log_variables: valid age_confidence/loss mean: 0.069803 2020-02-01 19:53:35.472: INFO @log_variables: valid age_confidence/accuracy mean: 0.563070 2020-02-01 19:53:35.472: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:53:35.479: INFO @metrics_hook: train age_mae: 6.073 +-0.035 (110592) 2020-02-01 19:53:35.486: INFO @metrics_hook: train gender_accuracy: 0.939 +-0.001 (110592) 2020-02-01 19:53:38.203: INFO @metrics_hook: valid age_mae: 6.364 +-0.090 (17639) 2020-02-01 19:53:38.205: INFO @metrics_hook: valid gender_accuracy: 0.916 +-0.004 (17639) 2020-02-01 19:53:39.662: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:53:39.662: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:53:39.662: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 19:53:39.663: INFO @evaluate_confidence: Average confidence of all samples 0.77 +- 0.26 2020-02-01 19:53:39.789: INFO @evaluate_confidence: Previous accuracy would be: 93.87 2020-02-01 19:53:39.790: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:53:39.852: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.18, 97.27, 97.38, 97.48, 97.57, 97.66, 97.73, 97.81, 97.88, 97.96, 98.02, 98.09, 98.18, 98.24, 98.29, 98.35, 98.41, 98.45, 98.51, 98.56, 98.6, 98.67, 98.71, 98.77, 98.8, 98.83, 98.88, 98.92, 98.97, 99.01, 99.04, 99.09, 99.14, 99.17, 99.2, 99.23, 99.26, 99.29, 99.32, 99.34, 99.37, 99.41, 99.43] 2020-02-01 19:53:39.853: INFO @evaluate_confidence: Dropped ratios are: [13.39, 13.95, 14.5, 15.03, 15.54, 16.08, 16.58, 17.1, 17.6, 18.15, 18.65, 19.15, 19.68, 20.21, 20.72, 21.24, 21.79, 22.32, 22.85, 23.37, 23.91, 24.49, 25.02, 25.57, 26.14, 26.68, 27.27, 27.85, 28.45, 29.07, 29.68, 30.32, 30.95, 31.57, 32.25, 32.97, 33.64, 34.34, 35.12, 35.88, 36.65, 37.43, 38.28] 2020-02-01 19:53:39.901: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:53:39.901: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 19:53:39.901: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:53:39.901: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:53:40.034: INFO @evaluate_confidence: Previous accuracy would be: 54.73 2020-02-01 19:53:40.034: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 19:53:40.053: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.54, 64.22, 64.91, 65.58, 66.32, 67.11, 67.85, 68.59, 69.42] 2020-02-01 19:53:40.053: INFO @evaluate_confidence: Dropped ratios are: [42.62, 45.55, 48.62, 51.61, 54.52, 57.29, 59.96, 62.5, 64.92] 2020-02-01 19:53:40.060: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:53:40.061: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.23 2020-02-01 19:53:40.061: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.20 2020-02-01 19:53:40.061: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.25 2020-02-01 19:53:40.158: INFO @evaluate_confidence: Previous accuracy would be: 91.64 2020-02-01 19:53:40.159: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 19:53:40.168: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.04, 96.19, 96.33, 96.42, 96.54, 96.63, 96.78, 96.87, 97.04, 97.15, 97.21, 97.32, 97.46, 97.56, 97.61, 97.69, 97.81, 97.88, 98.0, 98.07, 98.15, 98.22, 98.26, 98.36, 98.42, 98.48, 98.53, 98.58, 98.65, 98.69, 98.74, 98.76, 98.81, 98.85, 98.93, 99.02, 99.07, 99.13, 99.16, 99.18, 99.26] 2020-02-01 19:53:40.168: INFO @evaluate_confidence: Dropped ratios are: [13.71, 14.29, 14.77, 15.17, 15.73, 16.31, 16.89, 17.4, 17.89, 18.4, 18.88, 19.42, 19.87, 20.4, 20.89, 21.41, 21.99, 22.54, 23.08, 23.66, 24.2, 24.71, 25.22, 25.86, 26.52, 27.05, 27.67, 28.3, 28.84, 29.45, 30.14, 30.77, 31.4, 32.17, 32.95, 33.68, 34.56, 35.38, 36.2, 37.2, 38.04] 2020-02-01 19:53:40.175: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:53:40.176: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 19:53:40.176: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 19:53:40.176: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.11 2020-02-01 19:53:40.302: INFO @evaluate_confidence: Previous accuracy would be: 52.40 2020-02-01 19:53:40.303: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 19:53:40.304: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.74, 58.89, 59.54, 59.93] 2020-02-01 19:53:40.304: INFO @evaluate_confidence: Dropped ratios are: [44.69, 48.55, 52.59, 56.18] 2020-02-01 19:53:40.356: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:53:41.066: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:53:41.152: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:53:41.639: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:53:41.893: INFO @decay_lr : LR updated to `6.8320995e-05` 2020-02-01 19:53:41.894: INFO @log_profile : T train: 122.377487 2020-02-01 19:53:41.894: INFO @log_profile : T valid: 5.547307 2020-02-01 19:53:41.894: INFO @log_profile : T read data: 1.919043 2020-02-01 19:53:41.894: INFO @log_profile : T hooks: 7.046109 2020-02-01 19:53:41.894: INFO @main_loop : Epoch 76 done 2020-02-01 19:53:41.894: INFO @main_loop : Training epoch 77 2020-02-01 19:55:52.630: INFO @log_variables: train loss nanmean: 0.773603 2020-02-01 19:55:52.630: INFO @log_variables: train age_loss mean: 5.605082 2020-02-01 19:55:52.630: INFO @log_variables: train gender_loss mean: 0.151979 2020-02-01 19:55:52.630: INFO @log_variables: train age_mae mean: 6.082790 2020-02-01 19:55:52.630: INFO @log_variables: train gender_accuracy mean: 0.939160 2020-02-01 19:55:52.630: INFO @log_variables: train gender_confidence/loss nanmean: 0.057487 2020-02-01 19:55:52.631: INFO @log_variables: train gender_confidence/accuracy mean: 0.835919 2020-02-01 19:55:52.631: INFO @log_variables: train age_confidence/loss mean: 0.068401 2020-02-01 19:55:52.631: INFO @log_variables: train age_confidence/accuracy mean: 0.603070 2020-02-01 19:55:52.631: INFO @log_variables: valid loss nanmean: 0.837659 2020-02-01 19:55:52.631: INFO @log_variables: valid age_loss mean: 5.835745 2020-02-01 19:55:52.631: INFO @log_variables: valid gender_loss mean: 0.203066 2020-02-01 19:55:52.631: INFO @log_variables: valid age_mae mean: 6.314837 2020-02-01 19:55:52.631: INFO @log_variables: valid gender_accuracy mean: 0.917626 2020-02-01 19:55:52.631: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053476 2020-02-01 19:55:52.631: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864618 2020-02-01 19:55:52.631: INFO @log_variables: valid age_confidence/loss mean: 0.069056 2020-02-01 19:55:52.631: INFO @log_variables: valid age_confidence/accuracy mean: 0.559329 2020-02-01 19:55:52.631: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:55:52.639: INFO @metrics_hook: train age_mae: 6.083 +-0.035 (110372) 2020-02-01 19:55:52.646: INFO @metrics_hook: train gender_accuracy: 0.939 +-0.001 (110372) 2020-02-01 19:55:55.388: INFO @metrics_hook: valid age_mae: 6.315 +-0.089 (17639) 2020-02-01 19:55:55.390: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 19:55:57.004: INFO @decay_lr : LR updated to `6.797939e-05` 2020-02-01 19:55:57.005: INFO @log_profile : T train: 121.734104 2020-02-01 19:55:57.006: INFO @log_profile : T valid: 5.468106 2020-02-01 19:55:57.006: INFO @log_profile : T read data: 2.828683 2020-02-01 19:55:57.006: INFO @log_profile : T hooks: 5.004660 2020-02-01 19:55:57.006: INFO @main_loop : Epoch 77 done 2020-02-01 19:55:57.006: INFO @main_loop : Training epoch 78 2020-02-01 19:58:07.566: INFO @log_variables: train loss nanmean: 0.767368 2020-02-01 19:58:07.566: INFO @log_variables: train age_loss mean: 5.552222 2020-02-01 19:58:07.566: INFO @log_variables: train gender_loss mean: 0.150552 2020-02-01 19:58:07.566: INFO @log_variables: train age_mae mean: 6.029834 2020-02-01 19:58:07.566: INFO @log_variables: train gender_accuracy mean: 0.939785 2020-02-01 19:58:07.566: INFO @log_variables: train gender_confidence/loss nanmean: 0.057220 2020-02-01 19:58:07.566: INFO @log_variables: train gender_confidence/accuracy mean: 0.834931 2020-02-01 19:58:07.566: INFO @log_variables: train age_confidence/loss mean: 0.068535 2020-02-01 19:58:07.567: INFO @log_variables: train age_confidence/accuracy mean: 0.607020 2020-02-01 19:58:07.567: INFO @log_variables: valid loss nanmean: 0.838066 2020-02-01 19:58:07.567: INFO @log_variables: valid age_loss mean: 5.884946 2020-02-01 19:58:07.567: INFO @log_variables: valid gender_loss mean: 0.199337 2020-02-01 19:58:07.567: INFO @log_variables: valid age_mae mean: 6.364244 2020-02-01 19:58:07.567: INFO @log_variables: valid gender_accuracy mean: 0.918306 2020-02-01 19:58:07.567: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052171 2020-02-01 19:58:07.567: INFO @log_variables: valid gender_confidence/accuracy mean: 0.861897 2020-02-01 19:58:07.567: INFO @log_variables: valid age_confidence/loss mean: 0.069685 2020-02-01 19:58:07.567: INFO @log_variables: valid age_confidence/accuracy mean: 0.551732 2020-02-01 19:58:07.567: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 19:58:07.575: INFO @metrics_hook: train age_mae: 6.030 +-0.034 (110372) 2020-02-01 19:58:07.582: INFO @metrics_hook: train gender_accuracy: 0.940 +-0.001 (110372) 2020-02-01 19:58:10.310: INFO @metrics_hook: valid age_mae: 6.364 +-0.090 (17639) 2020-02-01 19:58:10.311: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 19:58:11.765: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:58:11.765: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 19:58:11.765: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 19:58:11.766: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 19:58:11.895: INFO @evaluate_confidence: Previous accuracy would be: 93.98 2020-02-01 19:58:11.895: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 19:58:11.959: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.29, 97.38, 97.47, 97.55, 97.65, 97.72, 97.79, 97.85, 97.92, 97.99, 98.05, 98.11, 98.16, 98.21, 98.26, 98.32, 98.37, 98.42, 98.48, 98.54, 98.6, 98.64, 98.7, 98.75, 98.79, 98.83, 98.86, 98.91, 98.95, 98.99, 99.03, 99.07, 99.11, 99.14, 99.18, 99.2, 99.24, 99.29, 99.32, 99.35, 99.38, 99.42, 99.45] 2020-02-01 19:58:11.960: INFO @evaluate_confidence: Dropped ratios are: [13.19, 13.75, 14.29, 14.77, 15.32, 15.86, 16.38, 16.93, 17.46, 18.01, 18.53, 19.05, 19.56, 20.09, 20.61, 21.11, 21.66, 22.17, 22.68, 23.24, 23.77, 24.3, 24.84, 25.42, 25.98, 26.53, 27.09, 27.71, 28.31, 28.93, 29.54, 30.17, 30.77, 31.44, 32.11, 32.77, 33.43, 34.14, 34.88, 35.61, 36.36, 37.2, 38.06] 2020-02-01 19:58:12.009: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:58:12.009: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 19:58:12.009: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 19:58:12.010: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 19:58:12.148: INFO @evaluate_confidence: Previous accuracy would be: 54.89 2020-02-01 19:58:12.148: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 19:58:12.167: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.66, 64.32, 65.09, 65.79, 66.65, 67.44, 68.13, 68.81, 69.67] 2020-02-01 19:58:12.168: INFO @evaluate_confidence: Dropped ratios are: [42.1, 45.3, 48.35, 51.44, 54.44, 57.23, 60.03, 62.61, 65.05] 2020-02-01 19:58:12.175: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 19:58:12.175: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.22 2020-02-01 19:58:12.176: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.21 2020-02-01 19:58:12.176: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.25 2020-02-01 19:58:12.279: INFO @evaluate_confidence: Previous accuracy would be: 91.83 2020-02-01 19:58:12.279: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 19:58:12.288: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.12, 96.21, 96.4, 96.52, 96.66, 96.8, 96.91, 96.97, 97.05, 97.14, 97.21, 97.33, 97.4, 97.48, 97.57, 97.63, 97.72, 97.81, 97.85, 97.92, 97.99, 98.08, 98.16, 98.25, 98.32, 98.41, 98.46, 98.57, 98.61, 98.66, 98.73, 98.77, 98.83, 98.84, 98.87, 98.97, 99.02, 99.06, 99.08, 99.12] 2020-02-01 19:58:12.288: INFO @evaluate_confidence: Dropped ratios are: [13.57, 14.03, 14.54, 14.92, 15.43, 15.93, 16.42, 16.95, 17.44, 18.02, 18.51, 19.01, 19.58, 20.06, 20.6, 21.13, 21.64, 22.13, 22.59, 23.09, 23.66, 24.18, 24.62, 25.15, 25.62, 26.24, 26.74, 27.34, 27.94, 28.62, 29.28, 29.84, 30.63, 31.3, 31.99, 32.83, 33.65, 34.47, 35.2, 36.01] 2020-02-01 19:58:12.296: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 19:58:12.296: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 19:58:12.296: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 19:58:12.296: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 19:58:12.425: INFO @evaluate_confidence: Previous accuracy would be: 52.43 2020-02-01 19:58:12.425: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 19:58:12.427: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.93, 59.32, 58.98, 59.3] 2020-02-01 19:58:12.427: INFO @evaluate_confidence: Dropped ratios are: [46.92, 51.16, 55.15, 59.15] 2020-02-01 19:58:12.482: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 19:58:13.209: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:58:13.295: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 19:58:13.775: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 19:58:14.019: INFO @decay_lr : LR updated to `6.763949e-05` 2020-02-01 19:58:14.020: INFO @log_profile : T train: 121.472871 2020-02-01 19:58:14.020: INFO @log_profile : T valid: 5.496528 2020-02-01 19:58:14.020: INFO @log_profile : T read data: 2.894017 2020-02-01 19:58:14.020: INFO @log_profile : T hooks: 7.072665 2020-02-01 19:58:14.020: INFO @main_loop : Epoch 78 done 2020-02-01 19:58:14.020: INFO @main_loop : Training epoch 79 2020-02-01 20:00:23.908: INFO @log_variables: train loss nanmean: 0.765546 2020-02-01 20:00:23.909: INFO @log_variables: train age_loss mean: 5.550139 2020-02-01 20:00:23.909: INFO @log_variables: train gender_loss mean: 0.148940 2020-02-01 20:00:23.909: INFO @log_variables: train age_mae mean: 6.027414 2020-02-01 20:00:23.909: INFO @log_variables: train gender_accuracy mean: 0.940547 2020-02-01 20:00:23.909: INFO @log_variables: train gender_confidence/loss nanmean: 0.057104 2020-02-01 20:00:23.909: INFO @log_variables: train gender_confidence/accuracy mean: 0.838542 2020-02-01 20:00:23.909: INFO @log_variables: train age_confidence/loss mean: 0.068484 2020-02-01 20:00:23.909: INFO @log_variables: train age_confidence/accuracy mean: 0.607214 2020-02-01 20:00:23.909: INFO @log_variables: valid loss nanmean: 0.849143 2020-02-01 20:00:23.909: INFO @log_variables: valid age_loss mean: 5.842500 2020-02-01 20:00:23.909: INFO @log_variables: valid gender_loss mean: 0.212534 2020-02-01 20:00:23.909: INFO @log_variables: valid age_mae mean: 6.321886 2020-02-01 20:00:23.909: INFO @log_variables: valid gender_accuracy mean: 0.911673 2020-02-01 20:00:23.909: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053836 2020-02-01 20:00:23.909: INFO @log_variables: valid gender_confidence/accuracy mean: 0.856851 2020-02-01 20:00:23.909: INFO @log_variables: valid age_confidence/loss mean: 0.070957 2020-02-01 20:00:23.909: INFO @log_variables: valid age_confidence/accuracy mean: 0.548444 2020-02-01 20:00:23.909: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:00:23.916: INFO @metrics_hook: train age_mae: 6.027 +-0.035 (110592) 2020-02-01 20:00:23.923: INFO @metrics_hook: train gender_accuracy: 0.941 +-0.001 (110592) 2020-02-01 20:00:26.652: INFO @metrics_hook: valid age_mae: 6.322 +-0.092 (17639) 2020-02-01 20:00:26.653: INFO @metrics_hook: valid gender_accuracy: 0.912 +-0.004 (17639) 2020-02-01 20:00:28.292: INFO @decay_lr : LR updated to `6.730129e-05` 2020-02-01 20:00:28.293: INFO @log_profile : T train: 121.889725 2020-02-01 20:00:28.293: INFO @log_profile : T valid: 5.432967 2020-02-01 20:00:28.293: INFO @log_profile : T read data: 1.895828 2020-02-01 20:00:28.293: INFO @log_profile : T hooks: 4.977246 2020-02-01 20:00:28.293: INFO @main_loop : Epoch 79 done 2020-02-01 20:00:28.293: INFO @main_loop : Training epoch 80 2020-02-01 20:02:40.484: INFO @log_variables: train loss nanmean: 0.767999 2020-02-01 20:02:40.484: INFO @log_variables: train age_loss mean: 5.555496 2020-02-01 20:02:40.484: INFO @log_variables: train gender_loss mean: 0.151128 2020-02-01 20:02:40.484: INFO @log_variables: train age_mae mean: 6.032790 2020-02-01 20:02:40.484: INFO @log_variables: train gender_accuracy mean: 0.939414 2020-02-01 20:02:40.485: INFO @log_variables: train gender_confidence/loss nanmean: 0.056881 2020-02-01 20:02:40.485: INFO @log_variables: train gender_confidence/accuracy mean: 0.837595 2020-02-01 20:02:40.485: INFO @log_variables: train age_confidence/loss mean: 0.068684 2020-02-01 20:02:40.485: INFO @log_variables: train age_confidence/accuracy mean: 0.605842 2020-02-01 20:02:40.485: INFO @log_variables: valid loss nanmean: 0.844621 2020-02-01 20:02:40.485: INFO @log_variables: valid age_loss mean: 5.844038 2020-02-01 20:02:40.485: INFO @log_variables: valid gender_loss mean: 0.208808 2020-02-01 20:02:40.485: INFO @log_variables: valid age_mae mean: 6.324234 2020-02-01 20:02:40.485: INFO @log_variables: valid gender_accuracy mean: 0.917286 2020-02-01 20:02:40.485: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053446 2020-02-01 20:02:40.485: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867226 2020-02-01 20:02:40.485: INFO @log_variables: valid age_confidence/loss mean: 0.070073 2020-02-01 20:02:40.485: INFO @log_variables: valid age_confidence/accuracy mean: 0.553433 2020-02-01 20:02:40.485: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:02:40.492: INFO @metrics_hook: train age_mae: 6.033 +-0.035 (110372) 2020-02-01 20:02:40.499: INFO @metrics_hook: train gender_accuracy: 0.939 +-0.001 (110372) 2020-02-01 20:02:43.247: INFO @metrics_hook: valid age_mae: 6.324 +-0.090 (17639) 2020-02-01 20:02:43.248: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 20:02:44.749: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:02:44.749: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 20:02:44.749: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 20:02:44.749: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:02:44.880: INFO @evaluate_confidence: Previous accuracy would be: 93.94 2020-02-01 20:02:44.880: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:02:44.942: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.24, 97.35, 97.43, 97.51, 97.61, 97.69, 97.76, 97.86, 97.93, 98.02, 98.07, 98.14, 98.21, 98.28, 98.32, 98.38, 98.44, 98.49, 98.54, 98.6, 98.64, 98.69, 98.75, 98.79, 98.82, 98.88, 98.92, 98.95, 98.99, 99.03, 99.08, 99.11, 99.15, 99.19, 99.21, 99.25, 99.28, 99.31, 99.34, 99.36, 99.39, 99.43, 99.46] 2020-02-01 20:02:44.942: INFO @evaluate_confidence: Dropped ratios are: [12.95, 13.54, 14.05, 14.62, 15.17, 15.69, 16.21, 16.76, 17.27, 17.82, 18.35, 18.89, 19.42, 19.94, 20.46, 20.98, 21.49, 22.02, 22.55, 23.1, 23.65, 24.2, 24.79, 25.34, 25.9, 26.45, 27.01, 27.63, 28.26, 28.86, 29.45, 30.09, 30.75, 31.41, 32.03, 32.71, 33.41, 34.1, 34.8, 35.54, 36.3, 37.08, 37.96] 2020-02-01 20:02:44.990: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:02:44.991: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:02:44.991: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 20:02:44.991: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 20:02:45.130: INFO @evaluate_confidence: Previous accuracy would be: 55.15 2020-02-01 20:02:45.130: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 20:02:45.149: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.89, 64.42, 65.1, 65.79, 66.59, 67.34, 68.02, 68.86, 69.77] 2020-02-01 20:02:45.149: INFO @evaluate_confidence: Dropped ratios are: [41.9, 44.86, 47.91, 50.88, 53.86, 56.63, 59.38, 62.01, 64.49] 2020-02-01 20:02:45.157: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:02:45.157: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.21 2020-02-01 20:02:45.157: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.22 2020-02-01 20:02:45.157: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.24 2020-02-01 20:02:45.261: INFO @evaluate_confidence: Previous accuracy would be: 91.73 2020-02-01 20:02:45.262: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:02:45.270: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.85, 95.93, 96.06, 96.16, 96.27, 96.39, 96.52, 96.67, 96.78, 96.9, 96.97, 97.07, 97.18, 97.26, 97.38, 97.45, 97.51, 97.54, 97.63, 97.69, 97.8, 97.87, 97.93, 97.98, 98.03, 98.12, 98.21, 98.3, 98.36, 98.43, 98.52, 98.62, 98.67, 98.76, 98.83, 98.85, 98.9, 98.96] 2020-02-01 20:02:45.270: INFO @evaluate_confidence: Dropped ratios are: [13.32, 13.73, 14.25, 14.74, 15.22, 15.61, 16.12, 16.72, 17.19, 17.65, 18.12, 18.67, 19.13, 19.55, 20.05, 20.56, 21.02, 21.45, 21.96, 22.57, 23.07, 23.52, 24.03, 24.51, 25.05, 25.67, 26.36, 26.92, 27.58, 28.3, 28.96, 29.68, 30.46, 31.32, 32.17, 33.03, 33.85, 34.92] 2020-02-01 20:02:45.278: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:02:45.278: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.50 +- 0.11 2020-02-01 20:02:45.278: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.10 2020-02-01 20:02:45.278: INFO @evaluate_confidence: Average confidence of all samples 0.48 +- 0.11 2020-02-01 20:02:45.405: INFO @evaluate_confidence: Previous accuracy would be: 52.87 2020-02-01 20:02:45.406: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49] 2020-02-01 20:02:45.407: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.69, 58.96, 59.46, 59.85] 2020-02-01 20:02:45.407: INFO @evaluate_confidence: Dropped ratios are: [44.19, 48.35, 52.46, 56.47] 2020-02-01 20:02:45.458: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:02:46.169: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:02:46.253: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:02:46.706: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:02:46.780: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:02:47.468: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:02:47.555: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:02:47.557: INFO @evaluate_gender-age_model: groups 0 4.047864 1 4.547611 2 5.672336 3 5.829509 4 6.704037 5 6.601779 6 6.768742 7 7.823483 Name: errors, dtype: float64 2020-02-01 20:02:47.558: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:02:48.006: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:02:48.068: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:02:48.069: INFO @evaluate_gender-age_model: groups 0 6.165972 1 5.656649 2 5.879318 3 5.678567 4 7.508334 5 5.440985 6 7.435961 7 11.926539 Name: errors, dtype: float64 2020-02-01 20:02:48.235: INFO @decay_lr : LR updated to `6.6964785e-05` 2020-02-01 20:02:48.237: INFO @log_profile : T train: 121.503749 2020-02-01 20:02:48.237: INFO @log_profile : T valid: 5.349727 2020-02-01 20:02:48.237: INFO @log_profile : T read data: 2.829020 2020-02-01 20:02:48.237: INFO @log_profile : T hooks: 10.183717 2020-02-01 20:02:48.237: INFO @main_loop : Epoch 80 done 2020-02-01 20:02:48.237: INFO @main_loop : Training epoch 81 2020-02-01 20:04:58.951: INFO @log_variables: train loss nanmean: 0.764672 2020-02-01 20:04:58.951: INFO @log_variables: train age_loss mean: 5.551399 2020-02-01 20:04:58.951: INFO @log_variables: train gender_loss mean: 0.147717 2020-02-01 20:04:58.951: INFO @log_variables: train age_mae mean: 6.028843 2020-02-01 20:04:58.951: INFO @log_variables: train gender_accuracy mean: 0.940528 2020-02-01 20:04:58.951: INFO @log_variables: train gender_confidence/loss nanmean: 0.057177 2020-02-01 20:04:58.951: INFO @log_variables: train gender_confidence/accuracy mean: 0.837712 2020-02-01 20:04:58.951: INFO @log_variables: train age_confidence/loss mean: 0.068534 2020-02-01 20:04:58.951: INFO @log_variables: train age_confidence/accuracy mean: 0.606241 2020-02-01 20:04:58.951: INFO @log_variables: valid loss nanmean: 0.854988 2020-02-01 20:04:58.951: INFO @log_variables: valid age_loss mean: 5.855727 2020-02-01 20:04:58.952: INFO @log_variables: valid gender_loss mean: 0.215485 2020-02-01 20:04:58.952: INFO @log_variables: valid age_mae mean: 6.335300 2020-02-01 20:04:58.952: INFO @log_variables: valid gender_accuracy mean: 0.911560 2020-02-01 20:04:58.952: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056850 2020-02-01 20:04:58.952: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862407 2020-02-01 20:04:58.952: INFO @log_variables: valid age_confidence/loss mean: 0.069905 2020-02-01 20:04:58.952: INFO @log_variables: valid age_confidence/accuracy mean: 0.554963 2020-02-01 20:04:58.952: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:04:58.959: INFO @metrics_hook: train age_mae: 6.029 +-0.035 (110372) 2020-02-01 20:04:58.966: INFO @metrics_hook: train gender_accuracy: 0.941 +-0.001 (110372) 2020-02-01 20:05:01.715: INFO @metrics_hook: valid age_mae: 6.335 +-0.091 (17639) 2020-02-01 20:05:01.716: INFO @metrics_hook: valid gender_accuracy: 0.912 +-0.004 (17639) 2020-02-01 20:05:03.381: INFO @decay_lr : LR updated to `6.662996e-05` 2020-02-01 20:05:03.382: INFO @log_profile : T train: 121.677370 2020-02-01 20:05:03.382: INFO @log_profile : T valid: 5.467005 2020-02-01 20:05:03.382: INFO @log_profile : T read data: 2.859625 2020-02-01 20:05:03.382: INFO @log_profile : T hooks: 5.065206 2020-02-01 20:05:03.382: INFO @main_loop : Epoch 81 done 2020-02-01 20:05:03.382: INFO @main_loop : Training epoch 82 2020-02-01 20:07:13.167: INFO @log_variables: train loss nanmean: 0.764271 2020-02-01 20:07:13.168: INFO @log_variables: train age_loss mean: 5.559519 2020-02-01 20:07:13.168: INFO @log_variables: train gender_loss mean: 0.147337 2020-02-01 20:07:13.168: INFO @log_variables: train age_mae mean: 6.036636 2020-02-01 20:07:13.168: INFO @log_variables: train gender_accuracy mean: 0.940493 2020-02-01 20:07:13.168: INFO @log_variables: train gender_confidence/loss nanmean: 0.056486 2020-02-01 20:07:13.168: INFO @log_variables: train gender_confidence/accuracy mean: 0.838135 2020-02-01 20:07:13.168: INFO @log_variables: train age_confidence/loss mean: 0.068432 2020-02-01 20:07:13.168: INFO @log_variables: train age_confidence/accuracy mean: 0.606744 2020-02-01 20:07:13.168: INFO @log_variables: valid loss nanmean: 0.834190 2020-02-01 20:07:13.168: INFO @log_variables: valid age_loss mean: 5.868156 2020-02-01 20:07:13.168: INFO @log_variables: valid gender_loss mean: 0.197104 2020-02-01 20:07:13.168: INFO @log_variables: valid age_mae mean: 6.348330 2020-02-01 20:07:13.168: INFO @log_variables: valid gender_accuracy mean: 0.920404 2020-02-01 20:07:13.168: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052238 2020-02-01 20:07:13.168: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864448 2020-02-01 20:07:13.168: INFO @log_variables: valid age_confidence/loss mean: 0.069298 2020-02-01 20:07:13.168: INFO @log_variables: valid age_confidence/accuracy mean: 0.568400 2020-02-01 20:07:13.168: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:07:13.176: INFO @metrics_hook: train age_mae: 6.037 +-0.034 (110592) 2020-02-01 20:07:13.183: INFO @metrics_hook: train gender_accuracy: 0.940 +-0.001 (110592) 2020-02-01 20:07:15.911: INFO @metrics_hook: valid age_mae: 6.348 +-0.091 (17639) 2020-02-01 20:07:15.912: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 20:07:17.387: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:07:17.387: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 20:07:17.387: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 20:07:17.387: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:07:17.513: INFO @evaluate_confidence: Previous accuracy would be: 94.05 2020-02-01 20:07:17.514: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:07:17.575: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.34, 97.45, 97.54, 97.62, 97.71, 97.78, 97.86, 97.95, 97.99, 98.04, 98.11, 98.18, 98.24, 98.3, 98.36, 98.41, 98.48, 98.54, 98.6, 98.65, 98.68, 98.73, 98.78, 98.81, 98.85, 98.89, 98.93, 98.98, 99.02, 99.07, 99.09, 99.13, 99.17, 99.2, 99.22, 99.25, 99.3, 99.35, 99.37, 99.4, 99.42, 99.45, 99.47] 2020-02-01 20:07:17.575: INFO @evaluate_confidence: Dropped ratios are: [13.07, 13.61, 14.16, 14.68, 15.18, 15.66, 16.18, 16.72, 17.21, 17.72, 18.27, 18.81, 19.3, 19.8, 20.32, 20.83, 21.35, 21.92, 22.43, 22.96, 23.5, 24.03, 24.59, 25.11, 25.68, 26.24, 26.81, 27.45, 28.01, 28.59, 29.2, 29.77, 30.38, 31.03, 31.69, 32.34, 33.01, 33.76, 34.49, 35.19, 35.94, 36.7, 37.52] 2020-02-01 20:07:17.624: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:07:17.624: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:07:17.624: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 20:07:17.624: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.14 2020-02-01 20:07:17.758: INFO @evaluate_confidence: Previous accuracy would be: 54.77 2020-02-01 20:07:17.758: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 20:07:17.777: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [63.66, 64.35, 64.96, 65.59, 66.29, 67.09, 67.79, 68.57, 69.33] 2020-02-01 20:07:17.777: INFO @evaluate_confidence: Dropped ratios are: [42.33, 45.4, 48.38, 51.28, 54.24, 57.02, 59.65, 62.27, 64.66] 2020-02-01 20:07:17.785: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:07:17.785: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.22 2020-02-01 20:07:17.785: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.22 2020-02-01 20:07:17.785: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 20:07:17.886: INFO @evaluate_confidence: Previous accuracy would be: 92.04 2020-02-01 20:07:17.886: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:07:17.895: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.1, 96.17, 96.27, 96.39, 96.49, 96.59, 96.71, 96.83, 96.85, 96.93, 96.99, 97.08, 97.17, 97.25, 97.42, 97.51, 97.58, 97.67, 97.76, 97.84, 97.92, 98.01, 98.08, 98.15, 98.27, 98.31, 98.37, 98.43, 98.53, 98.54, 98.63, 98.68, 98.74, 98.77, 98.8, 98.85, 98.89, 98.94, 99.02, 99.05, 99.09] 2020-02-01 20:07:17.895: INFO @evaluate_confidence: Dropped ratios are: [13.22, 13.63, 14.03, 14.47, 14.88, 15.31, 15.76, 16.21, 16.5, 16.93, 17.4, 17.82, 18.26, 18.65, 19.28, 19.74, 20.24, 20.68, 21.15, 21.68, 22.17, 22.65, 23.09, 23.56, 24.11, 24.56, 25.05, 25.58, 26.17, 26.76, 27.33, 28.05, 28.74, 29.39, 30.1, 30.8, 31.53, 32.39, 33.18, 34.08, 34.88] 2020-02-01 20:07:17.902: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:07:17.902: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:07:17.903: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 20:07:17.903: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 20:07:18.025: INFO @evaluate_confidence: Previous accuracy would be: 52.58 2020-02-01 20:07:18.025: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51] 2020-02-01 20:07:18.027: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.69, 60.04, 60.39, 60.42, 60.85] 2020-02-01 20:07:18.027: INFO @evaluate_confidence: Dropped ratios are: [42.21, 46.08, 50.65, 54.83, 58.88] 2020-02-01 20:07:18.078: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:07:18.774: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:07:18.857: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:07:19.302: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:07:19.534: INFO @decay_lr : LR updated to `6.629681e-05` 2020-02-01 20:07:19.535: INFO @log_profile : T train: 121.709327 2020-02-01 20:07:19.535: INFO @log_profile : T valid: 5.440293 2020-02-01 20:07:19.535: INFO @log_profile : T read data: 1.942920 2020-02-01 20:07:19.535: INFO @log_profile : T hooks: 6.983618 2020-02-01 20:07:19.536: INFO @main_loop : Epoch 82 done 2020-02-01 20:07:19.536: INFO @main_loop : Training epoch 83 2020-02-01 20:09:30.289: INFO @log_variables: train loss nanmean: 0.755723 2020-02-01 20:09:30.289: INFO @log_variables: train age_loss mean: 5.506706 2020-02-01 20:09:30.289: INFO @log_variables: train gender_loss mean: 0.143488 2020-02-01 20:09:30.289: INFO @log_variables: train age_mae mean: 5.983979 2020-02-01 20:09:30.289: INFO @log_variables: train gender_accuracy mean: 0.942558 2020-02-01 20:09:30.289: INFO @log_variables: train gender_confidence/loss nanmean: 0.055719 2020-02-01 20:09:30.289: INFO @log_variables: train gender_confidence/accuracy mean: 0.840874 2020-02-01 20:09:30.289: INFO @log_variables: train age_confidence/loss mean: 0.068951 2020-02-01 20:09:30.289: INFO @log_variables: train age_confidence/accuracy mean: 0.605733 2020-02-01 20:09:30.289: INFO @log_variables: valid loss nanmean: 0.873997 2020-02-01 20:09:30.289: INFO @log_variables: valid age_loss mean: 5.877340 2020-02-01 20:09:30.289: INFO @log_variables: valid gender_loss mean: 0.238447 2020-02-01 20:09:30.289: INFO @log_variables: valid age_mae mean: 6.357083 2020-02-01 20:09:30.290: INFO @log_variables: valid gender_accuracy mean: 0.905323 2020-02-01 20:09:30.290: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053233 2020-02-01 20:09:30.290: INFO @log_variables: valid gender_confidence/accuracy mean: 0.852373 2020-02-01 20:09:30.290: INFO @log_variables: valid age_confidence/loss mean: 0.069690 2020-02-01 20:09:30.290: INFO @log_variables: valid age_confidence/accuracy mean: 0.555757 2020-02-01 20:09:30.290: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:09:30.297: INFO @metrics_hook: train age_mae: 5.984 +-0.034 (110372) 2020-02-01 20:09:30.304: INFO @metrics_hook: train gender_accuracy: 0.943 +-0.001 (110372) 2020-02-01 20:09:33.027: INFO @metrics_hook: valid age_mae: 6.357 +-0.090 (17639) 2020-02-01 20:09:33.028: INFO @metrics_hook: valid gender_accuracy: 0.905 +-0.004 (17639) 2020-02-01 20:09:34.656: INFO @decay_lr : LR updated to `6.596532e-05` 2020-02-01 20:09:34.657: INFO @log_profile : T train: 121.714592 2020-02-01 20:09:34.657: INFO @log_profile : T valid: 5.480177 2020-02-01 20:09:34.657: INFO @log_profile : T read data: 2.854161 2020-02-01 20:09:34.657: INFO @log_profile : T hooks: 4.993935 2020-02-01 20:09:34.657: INFO @main_loop : Epoch 83 done 2020-02-01 20:09:34.657: INFO @main_loop : Training epoch 84 2020-02-01 20:11:45.142: INFO @log_variables: train loss nanmean: 0.757165 2020-02-01 20:11:45.142: INFO @log_variables: train age_loss mean: 5.492020 2020-02-01 20:11:45.142: INFO @log_variables: train gender_loss mean: 0.145792 2020-02-01 20:11:45.142: INFO @log_variables: train age_mae mean: 5.968944 2020-02-01 20:11:45.142: INFO @log_variables: train gender_accuracy mean: 0.941534 2020-02-01 20:11:45.142: INFO @log_variables: train gender_confidence/loss nanmean: 0.056511 2020-02-01 20:11:45.142: INFO @log_variables: train gender_confidence/accuracy mean: 0.837767 2020-02-01 20:11:45.142: INFO @log_variables: train age_confidence/loss mean: 0.068842 2020-02-01 20:11:45.142: INFO @log_variables: train age_confidence/accuracy mean: 0.607862 2020-02-01 20:11:45.142: INFO @log_variables: valid loss nanmean: 0.849939 2020-02-01 20:11:45.142: INFO @log_variables: valid age_loss mean: 5.864813 2020-02-01 20:11:45.142: INFO @log_variables: valid gender_loss mean: 0.209688 2020-02-01 20:11:45.142: INFO @log_variables: valid age_mae mean: 6.343850 2020-02-01 20:11:45.142: INFO @log_variables: valid gender_accuracy mean: 0.915188 2020-02-01 20:11:45.142: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056439 2020-02-01 20:11:45.143: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870287 2020-02-01 20:11:45.143: INFO @log_variables: valid age_confidence/loss mean: 0.069710 2020-02-01 20:11:45.143: INFO @log_variables: valid age_confidence/accuracy mean: 0.563467 2020-02-01 20:11:45.143: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:11:45.150: INFO @metrics_hook: train age_mae: 5.969 +-0.034 (110372) 2020-02-01 20:11:45.157: INFO @metrics_hook: train gender_accuracy: 0.942 +-0.001 (110372) 2020-02-01 20:11:47.941: INFO @metrics_hook: valid age_mae: 6.344 +-0.090 (17639) 2020-02-01 20:11:47.943: INFO @metrics_hook: valid gender_accuracy: 0.915 +-0.004 (17639) 2020-02-01 20:11:54.510: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:11:54.511: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.80 +- 0.24 2020-02-01 20:11:54.511: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 20:11:54.511: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:11:54.643: INFO @evaluate_confidence: Previous accuracy would be: 94.15 2020-02-01 20:11:54.644: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:11:54.703: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.4, 97.47, 97.57, 97.66, 97.74, 97.82, 97.89, 97.96, 98.04, 98.11, 98.18, 98.25, 98.32, 98.38, 98.43, 98.51, 98.56, 98.61, 98.67, 98.71, 98.76, 98.81, 98.86, 98.89, 98.92, 98.96, 98.97, 99.02, 99.06, 99.09, 99.14, 99.17, 99.21, 99.24, 99.27, 99.3, 99.33, 99.34, 99.36, 99.39, 99.43, 99.45, 99.49] 2020-02-01 20:11:54.703: INFO @evaluate_confidence: Dropped ratios are: [13.25, 13.8, 14.35, 14.83, 15.37, 15.87, 16.37, 16.85, 17.36, 17.88, 18.38, 18.86, 19.36, 19.88, 20.33, 20.85, 21.36, 21.86, 22.39, 22.92, 23.5, 24.03, 24.55, 25.05, 25.6, 26.14, 26.69, 27.25, 27.83, 28.42, 29.05, 29.62, 30.23, 30.83, 31.42, 32.09, 32.74, 33.45, 34.12, 34.85, 35.59, 36.39, 37.24] 2020-02-01 20:11:54.752: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:11:54.752: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:11:54.753: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.12 2020-02-01 20:11:54.753: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:11:54.890: INFO @evaluate_confidence: Previous accuracy would be: 55.50 2020-02-01 20:11:54.891: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 20:11:54.910: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.15, 64.74, 65.31, 65.98, 66.66, 67.43, 68.25, 69.01, 69.92] 2020-02-01 20:11:54.910: INFO @evaluate_confidence: Dropped ratios are: [40.79, 43.76, 46.81, 49.72, 52.58, 55.49, 58.24, 60.98, 63.54] 2020-02-01 20:11:54.917: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:11:54.917: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.21 2020-02-01 20:11:54.918: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.22 2020-02-01 20:11:54.918: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 20:11:55.022: INFO @evaluate_confidence: Previous accuracy would be: 91.52 2020-02-01 20:11:55.023: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:11:55.031: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.69, 95.81, 95.94, 96.11, 96.22, 96.35, 96.46, 96.6, 96.66, 96.77, 96.87, 96.94, 97.08, 97.19, 97.3, 97.38, 97.44, 97.5, 97.55, 97.63, 97.68, 97.76, 97.88, 97.97, 98.05, 98.12, 98.15, 98.24, 98.34, 98.42, 98.5, 98.55, 98.63, 98.7, 98.78, 98.86, 98.9, 98.96, 99.01] 2020-02-01 20:11:55.031: INFO @evaluate_confidence: Dropped ratios are: [13.31, 13.76, 14.23, 14.83, 15.28, 15.75, 16.26, 16.82, 17.26, 17.76, 18.18, 18.65, 19.17, 19.76, 20.2, 20.78, 21.32, 21.86, 22.39, 22.97, 23.53, 24.14, 24.72, 25.32, 25.9, 26.46, 27.07, 27.7, 28.4, 29.09, 29.76, 30.41, 31.08, 31.86, 32.69, 33.6, 34.45, 35.31, 36.28] 2020-02-01 20:11:55.039: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:11:55.039: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:11:55.039: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 20:11:55.040: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 20:11:55.172: INFO @evaluate_confidence: Previous accuracy would be: 52.51 2020-02-01 20:11:55.172: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:11:55.174: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.09, 59.52, 59.88, 59.94] 2020-02-01 20:11:55.174: INFO @evaluate_confidence: Dropped ratios are: [46.5, 50.89, 55.14, 59.07] 2020-02-01 20:11:55.224: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:11:55.918: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:11:56.003: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:11:56.455: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:11:56.700: INFO @decay_lr : LR updated to `6.56355e-05` 2020-02-01 20:11:56.701: INFO @log_profile : T train: 121.480317 2020-02-01 20:11:56.701: INFO @log_profile : T valid: 5.386079 2020-02-01 20:11:56.701: INFO @log_profile : T read data: 2.924201 2020-02-01 20:11:56.702: INFO @log_profile : T hooks: 12.175669 2020-02-01 20:11:56.702: INFO @main_loop : Epoch 84 done 2020-02-01 20:11:56.702: INFO @main_loop : Training epoch 85 2020-02-01 20:14:06.619: INFO @log_variables: train loss nanmean: 0.756957 2020-02-01 20:14:06.619: INFO @log_variables: train age_loss mean: 5.488076 2020-02-01 20:14:06.619: INFO @log_variables: train gender_loss mean: 0.145220 2020-02-01 20:14:06.619: INFO @log_variables: train age_mae mean: 5.964834 2020-02-01 20:14:06.619: INFO @log_variables: train gender_accuracy mean: 0.942256 2020-02-01 20:14:06.619: INFO @log_variables: train gender_confidence/loss nanmean: 0.057202 2020-02-01 20:14:06.620: INFO @log_variables: train gender_confidence/accuracy mean: 0.838659 2020-02-01 20:14:06.620: INFO @log_variables: train age_confidence/loss mean: 0.068821 2020-02-01 20:14:06.620: INFO @log_variables: train age_confidence/accuracy mean: 0.606762 2020-02-01 20:14:06.620: INFO @log_variables: valid loss nanmean: 0.837355 2020-02-01 20:14:06.620: INFO @log_variables: valid age_loss mean: 5.855126 2020-02-01 20:14:06.620: INFO @log_variables: valid gender_loss mean: 0.200096 2020-02-01 20:14:06.620: INFO @log_variables: valid age_mae mean: 6.335449 2020-02-01 20:14:06.620: INFO @log_variables: valid gender_accuracy mean: 0.920177 2020-02-01 20:14:06.620: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053629 2020-02-01 20:14:06.620: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871535 2020-02-01 20:14:06.620: INFO @log_variables: valid age_confidence/loss mean: 0.069537 2020-02-01 20:14:06.620: INFO @log_variables: valid age_confidence/accuracy mean: 0.555814 2020-02-01 20:14:06.620: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:14:06.627: INFO @metrics_hook: train age_mae: 5.965 +-0.034 (110592) 2020-02-01 20:14:06.634: INFO @metrics_hook: train gender_accuracy: 0.942 +-0.001 (110592) 2020-02-01 20:14:09.368: INFO @metrics_hook: valid age_mae: 6.335 +-0.089 (17639) 2020-02-01 20:14:09.369: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 20:14:11.057: INFO @decay_lr : LR updated to `6.530732e-05` 2020-02-01 20:14:11.058: INFO @log_profile : T train: 121.906072 2020-02-01 20:14:11.058: INFO @log_profile : T valid: 5.424893 2020-02-01 20:14:11.058: INFO @log_profile : T read data: 1.881104 2020-02-01 20:14:11.058: INFO @log_profile : T hooks: 5.066451 2020-02-01 20:14:11.058: INFO @main_loop : Epoch 85 done 2020-02-01 20:14:11.058: INFO @main_loop : Training epoch 86 2020-02-01 20:16:30.660: INFO @log_variables: train loss nanmean: 0.752445 2020-02-01 20:16:30.660: INFO @log_variables: train age_loss mean: 5.464710 2020-02-01 20:16:30.660: INFO @log_variables: train gender_loss mean: 0.143170 2020-02-01 20:16:30.660: INFO @log_variables: train age_mae mean: 5.941633 2020-02-01 20:16:30.660: INFO @log_variables: train gender_accuracy mean: 0.943011 2020-02-01 20:16:30.660: INFO @log_variables: train gender_confidence/loss nanmean: 0.056370 2020-02-01 20:16:30.660: INFO @log_variables: train gender_confidence/accuracy mean: 0.838655 2020-02-01 20:16:30.660: INFO @log_variables: train age_confidence/loss mean: 0.069129 2020-02-01 20:16:30.660: INFO @log_variables: train age_confidence/accuracy mean: 0.604220 2020-02-01 20:16:30.660: INFO @log_variables: valid loss nanmean: 0.831523 2020-02-01 20:16:30.660: INFO @log_variables: valid age_loss mean: 5.905344 2020-02-01 20:16:30.660: INFO @log_variables: valid gender_loss mean: 0.190985 2020-02-01 20:16:30.660: INFO @log_variables: valid age_mae mean: 6.386234 2020-02-01 20:16:30.660: INFO @log_variables: valid gender_accuracy mean: 0.922388 2020-02-01 20:16:30.661: INFO @log_variables: valid gender_confidence/loss nanmean: 0.051990 2020-02-01 20:16:30.661: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873632 2020-02-01 20:16:30.661: INFO @log_variables: valid age_confidence/loss mean: 0.069061 2020-02-01 20:16:30.661: INFO @log_variables: valid age_confidence/accuracy mean: 0.567549 2020-02-01 20:16:30.661: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:16:30.668: INFO @metrics_hook: train age_mae: 5.942 +-0.034 (110372) 2020-02-01 20:16:30.675: INFO @metrics_hook: train gender_accuracy: 0.943 +-0.001 (110372) 2020-02-01 20:16:33.448: INFO @metrics_hook: valid age_mae: 6.386 +-0.089 (17639) 2020-02-01 20:16:33.449: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 20:16:34.925: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:16:34.925: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:16:34.925: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 20:16:34.926: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:16:35.057: INFO @evaluate_confidence: Previous accuracy would be: 94.30 2020-02-01 20:16:35.057: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 20:16:35.122: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.42, 97.51, 97.58, 97.67, 97.75, 97.82, 97.9, 97.96, 98.05, 98.14, 98.2, 98.26, 98.34, 98.4, 98.47, 98.52, 98.57, 98.63, 98.67, 98.72, 98.76, 98.8, 98.84, 98.89, 98.93, 98.96, 99.0, 99.03, 99.06, 99.1, 99.14, 99.17, 99.2, 99.25, 99.27, 99.3, 99.33, 99.37, 99.4, 99.43, 99.46, 99.48, 99.51, 99.54] 2020-02-01 20:16:35.122: INFO @evaluate_confidence: Dropped ratios are: [12.97, 13.49, 13.99, 14.51, 15.0, 15.51, 16.06, 16.55, 17.07, 17.56, 18.09, 18.64, 19.16, 19.65, 20.13, 20.6, 21.07, 21.57, 22.1, 22.58, 23.11, 23.63, 24.19, 24.74, 25.28, 25.82, 26.36, 26.92, 27.47, 28.06, 28.69, 29.23, 29.88, 30.48, 31.1, 31.75, 32.43, 33.12, 33.82, 34.54, 35.29, 36.13, 36.92, 37.75] 2020-02-01 20:16:35.172: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:16:35.172: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:16:35.172: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:16:35.173: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:16:35.313: INFO @evaluate_confidence: Previous accuracy would be: 55.76 2020-02-01 20:16:35.313: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:16:35.328: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.54, 65.12, 65.82, 66.66, 67.36, 68.12, 69.04] 2020-02-01 20:16:35.328: INFO @evaluate_confidence: Dropped ratios are: [43.41, 46.51, 49.55, 52.48, 55.47, 58.33, 60.99] 2020-02-01 20:16:35.336: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:16:35.336: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.21 2020-02-01 20:16:35.336: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.22 2020-02-01 20:16:35.336: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 20:16:35.441: INFO @evaluate_confidence: Previous accuracy would be: 92.24 2020-02-01 20:16:35.441: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:16:35.450: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.25, 96.36, 96.44, 96.61, 96.72, 96.81, 96.91, 96.96, 97.06, 97.19, 97.28, 97.37, 97.42, 97.47, 97.52, 97.6, 97.7, 97.76, 97.88, 97.94, 98.04, 98.13, 98.2, 98.27, 98.32, 98.41, 98.44, 98.52, 98.58, 98.68, 98.72, 98.81, 98.84, 98.87, 98.93, 98.98, 99.0, 99.07, 99.15, 99.22, 99.27] 2020-02-01 20:16:35.450: INFO @evaluate_confidence: Dropped ratios are: [12.76, 13.14, 13.57, 14.05, 14.46, 14.97, 15.39, 15.83, 16.25, 16.84, 17.35, 17.79, 18.25, 18.69, 19.15, 19.67, 20.18, 20.65, 21.18, 21.55, 22.09, 22.57, 23.13, 23.7, 24.25, 24.74, 25.34, 25.9, 26.5, 27.09, 27.6, 28.33, 29.0, 29.66, 30.34, 30.95, 31.73, 32.54, 33.41, 34.28, 35.22] 2020-02-01 20:16:35.458: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:16:35.458: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.10 2020-02-01 20:16:35.458: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.09 2020-02-01 20:16:35.458: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 20:16:35.587: INFO @evaluate_confidence: Previous accuracy would be: 51.89 2020-02-01 20:16:35.587: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:16:35.589: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.77, 59.18, 59.29, 59.49] 2020-02-01 20:16:35.589: INFO @evaluate_confidence: Dropped ratios are: [43.93, 48.72, 53.45, 58.0] 2020-02-01 20:16:35.642: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:16:36.380: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:16:36.465: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:16:36.936: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:16:37.176: INFO @decay_lr : LR updated to `6.4980784e-05` 2020-02-01 20:16:37.177: INFO @log_profile : T train: 129.791169 2020-02-01 20:16:37.177: INFO @log_profile : T valid: 6.346536 2020-02-01 20:16:37.177: INFO @log_profile : T read data: 2.784673 2020-02-01 20:16:37.178: INFO @log_profile : T hooks: 7.120470 2020-02-01 20:16:37.178: INFO @main_loop : Epoch 86 done 2020-02-01 20:16:37.178: INFO @main_loop : Training epoch 87 2020-02-01 20:18:57.809: INFO @log_variables: train loss nanmean: 0.749288 2020-02-01 20:18:57.810: INFO @log_variables: train age_loss mean: 5.442977 2020-02-01 20:18:57.810: INFO @log_variables: train gender_loss mean: 0.143121 2020-02-01 20:18:57.810: INFO @log_variables: train age_mae mean: 5.919916 2020-02-01 20:18:57.810: INFO @log_variables: train gender_accuracy mean: 0.942395 2020-02-01 20:18:57.810: INFO @log_variables: train gender_confidence/loss nanmean: 0.055528 2020-02-01 20:18:57.810: INFO @log_variables: train gender_confidence/accuracy mean: 0.840938 2020-02-01 20:18:57.810: INFO @log_variables: train age_confidence/loss mean: 0.068833 2020-02-01 20:18:57.810: INFO @log_variables: train age_confidence/accuracy mean: 0.609620 2020-02-01 20:18:57.810: INFO @log_variables: valid loss nanmean: 0.830855 2020-02-01 20:18:57.810: INFO @log_variables: valid age_loss mean: 5.884440 2020-02-01 20:18:57.810: INFO @log_variables: valid gender_loss mean: 0.190958 2020-02-01 20:18:57.810: INFO @log_variables: valid age_mae mean: 6.364703 2020-02-01 20:18:57.810: INFO @log_variables: valid gender_accuracy mean: 0.922331 2020-02-01 20:18:57.810: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052821 2020-02-01 20:18:57.811: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872782 2020-02-01 20:18:57.811: INFO @log_variables: valid age_confidence/loss mean: 0.069486 2020-02-01 20:18:57.811: INFO @log_variables: valid age_confidence/accuracy mean: 0.555644 2020-02-01 20:18:57.811: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:18:57.818: INFO @metrics_hook: train age_mae: 5.920 +-0.034 (110372) 2020-02-01 20:18:57.826: INFO @metrics_hook: train gender_accuracy: 0.942 +-0.001 (110372) 2020-02-01 20:19:00.632: INFO @metrics_hook: valid age_mae: 6.365 +-0.090 (17639) 2020-02-01 20:19:00.633: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 20:19:02.348: INFO @decay_lr : LR updated to `6.4655884e-05` 2020-02-01 20:19:02.349: INFO @log_profile : T train: 130.858614 2020-02-01 20:19:02.349: INFO @log_profile : T valid: 6.320293 2020-02-01 20:19:02.349: INFO @log_profile : T read data: 2.764280 2020-02-01 20:19:02.349: INFO @log_profile : T hooks: 5.152661 2020-02-01 20:19:02.349: INFO @main_loop : Epoch 87 done 2020-02-01 20:19:02.349: INFO @main_loop : Training epoch 88 2020-02-01 20:21:21.584: INFO @log_variables: train loss nanmean: 0.746920 2020-02-01 20:21:21.584: INFO @log_variables: train age_loss mean: 5.403907 2020-02-01 20:21:21.584: INFO @log_variables: train gender_loss mean: 0.144110 2020-02-01 20:21:21.584: INFO @log_variables: train age_mae mean: 5.880701 2020-02-01 20:21:21.584: INFO @log_variables: train gender_accuracy mean: 0.941840 2020-02-01 20:21:21.584: INFO @log_variables: train gender_confidence/loss nanmean: 0.055529 2020-02-01 20:21:21.584: INFO @log_variables: train gender_confidence/accuracy mean: 0.839735 2020-02-01 20:21:21.584: INFO @log_variables: train age_confidence/loss mean: 0.069118 2020-02-01 20:21:21.584: INFO @log_variables: train age_confidence/accuracy mean: 0.607639 2020-02-01 20:21:21.584: INFO @log_variables: valid loss nanmean: 0.835913 2020-02-01 20:21:21.585: INFO @log_variables: valid age_loss mean: 5.860844 2020-02-01 20:21:21.585: INFO @log_variables: valid gender_loss mean: 0.197584 2020-02-01 20:21:21.585: INFO @log_variables: valid age_mae mean: 6.341921 2020-02-01 20:21:21.585: INFO @log_variables: valid gender_accuracy mean: 0.920460 2020-02-01 20:21:21.585: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054022 2020-02-01 20:21:21.585: INFO @log_variables: valid gender_confidence/accuracy mean: 0.879075 2020-02-01 20:21:21.585: INFO @log_variables: valid age_confidence/loss mean: 0.069466 2020-02-01 20:21:21.585: INFO @log_variables: valid age_confidence/accuracy mean: 0.552015 2020-02-01 20:21:21.585: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:21:21.592: INFO @metrics_hook: train age_mae: 5.881 +-0.034 (110592) 2020-02-01 20:21:21.599: INFO @metrics_hook: train gender_accuracy: 0.942 +-0.001 (110592) 2020-02-01 20:21:24.403: INFO @metrics_hook: valid age_mae: 6.342 +-0.089 (17639) 2020-02-01 20:21:24.404: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 20:21:25.874: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:21:25.875: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:21:25.875: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 20:21:25.875: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:21:26.006: INFO @evaluate_confidence: Previous accuracy would be: 94.18 2020-02-01 20:21:26.006: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:21:26.069: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.44, 97.53, 97.6, 97.67, 97.75, 97.84, 97.92, 98.0, 98.08, 98.15, 98.21, 98.27, 98.32, 98.36, 98.42, 98.47, 98.53, 98.57, 98.62, 98.67, 98.72, 98.76, 98.79, 98.84, 98.87, 98.92, 98.95, 98.98, 99.02, 99.06, 99.09, 99.12, 99.14, 99.19, 99.21, 99.25, 99.28, 99.31, 99.34, 99.36, 99.39, 99.42, 99.44, 99.46] 2020-02-01 20:21:26.069: INFO @evaluate_confidence: Dropped ratios are: [12.79, 13.29, 13.78, 14.28, 14.8, 15.3, 15.83, 16.33, 16.82, 17.3, 17.76, 18.25, 18.71, 19.19, 19.67, 20.15, 20.67, 21.15, 21.64, 22.11, 22.59, 23.11, 23.59, 24.08, 24.58, 25.1, 25.66, 26.18, 26.74, 27.25, 27.83, 28.38, 28.94, 29.56, 30.19, 30.8, 31.46, 32.15, 32.82, 33.49, 34.27, 34.99, 35.74, 36.53] 2020-02-01 20:21:26.120: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:21:26.120: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:21:26.120: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:21:26.121: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:21:26.257: INFO @evaluate_confidence: Previous accuracy would be: 55.96 2020-02-01 20:21:26.258: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:21:26.273: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.95, 65.6, 66.32, 66.92, 67.62, 68.36, 69.14] 2020-02-01 20:21:26.273: INFO @evaluate_confidence: Dropped ratios are: [43.45, 46.39, 49.31, 52.26, 54.99, 57.8, 60.45] 2020-02-01 20:21:26.280: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:21:26.281: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.21 2020-02-01 20:21:26.281: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.22 2020-02-01 20:21:26.281: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 20:21:26.385: INFO @evaluate_confidence: Previous accuracy would be: 92.05 2020-02-01 20:21:26.386: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 20:21:26.394: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.94, 96.12, 96.2, 96.32, 96.45, 96.54, 96.66, 96.78, 96.87, 96.99, 97.11, 97.21, 97.35, 97.42, 97.52, 97.62, 97.71, 97.78, 97.92, 97.98, 98.05, 98.11, 98.18, 98.25, 98.3, 98.35, 98.43, 98.51, 98.54, 98.58, 98.66, 98.72, 98.83, 98.9, 98.96, 99.01, 99.03, 99.06, 99.12, 99.16, 99.22] 2020-02-01 20:21:26.395: INFO @evaluate_confidence: Dropped ratios are: [12.29, 12.73, 13.07, 13.52, 13.94, 14.44, 14.89, 15.4, 15.88, 16.38, 16.91, 17.29, 17.77, 18.19, 18.57, 19.13, 19.57, 20.0, 20.53, 20.96, 21.54, 21.99, 22.47, 23.04, 23.61, 24.09, 24.66, 25.21, 25.79, 26.45, 27.16, 27.75, 28.57, 29.2, 30.02, 30.74, 31.51, 32.38, 33.26, 34.23, 35.23] 2020-02-01 20:21:26.402: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:21:26.402: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 20:21:26.403: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.09 2020-02-01 20:21:26.403: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 20:21:26.530: INFO @evaluate_confidence: Previous accuracy would be: 51.90 2020-02-01 20:21:26.530: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 20:21:26.532: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.69, 57.98, 58.48, 58.73] 2020-02-01 20:21:26.532: INFO @evaluate_confidence: Dropped ratios are: [45.61, 50.12, 54.72, 59.36] 2020-02-01 20:21:26.583: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:21:27.287: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:21:27.371: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:21:27.839: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:21:28.083: INFO @decay_lr : LR updated to `6.4332606e-05` 2020-02-01 20:21:28.085: INFO @log_profile : T train: 130.414762 2020-02-01 20:21:28.085: INFO @log_profile : T valid: 6.254787 2020-02-01 20:21:28.086: INFO @log_profile : T read data: 1.851690 2020-02-01 20:21:28.086: INFO @log_profile : T hooks: 7.137921 2020-02-01 20:21:28.086: INFO @main_loop : Epoch 88 done 2020-02-01 20:21:28.086: INFO @main_loop : Training epoch 89 2020-02-01 20:23:48.152: INFO @log_variables: train loss nanmean: 0.749358 2020-02-01 20:23:48.153: INFO @log_variables: train age_loss mean: 5.440896 2020-02-01 20:23:48.153: INFO @log_variables: train gender_loss mean: 0.142953 2020-02-01 20:23:48.153: INFO @log_variables: train age_mae mean: 5.918079 2020-02-01 20:23:48.153: INFO @log_variables: train gender_accuracy mean: 0.941833 2020-02-01 20:23:48.153: INFO @log_variables: train gender_confidence/loss nanmean: 0.055706 2020-02-01 20:23:48.153: INFO @log_variables: train gender_confidence/accuracy mean: 0.841636 2020-02-01 20:23:48.153: INFO @log_variables: train age_confidence/loss mean: 0.069067 2020-02-01 20:23:48.153: INFO @log_variables: train age_confidence/accuracy mean: 0.607554 2020-02-01 20:23:48.153: INFO @log_variables: valid loss nanmean: 0.838436 2020-02-01 20:23:48.153: INFO @log_variables: valid age_loss mean: 5.906892 2020-02-01 20:23:48.153: INFO @log_variables: valid gender_loss mean: 0.195305 2020-02-01 20:23:48.153: INFO @log_variables: valid age_mae mean: 6.387342 2020-02-01 20:23:48.153: INFO @log_variables: valid gender_accuracy mean: 0.920460 2020-02-01 20:23:48.153: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054607 2020-02-01 20:23:48.153: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866943 2020-02-01 20:23:48.153: INFO @log_variables: valid age_confidence/loss mean: 0.069288 2020-02-01 20:23:48.153: INFO @log_variables: valid age_confidence/accuracy mean: 0.554850 2020-02-01 20:23:48.154: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:23:48.161: INFO @metrics_hook: train age_mae: 5.918 +-0.034 (110372) 2020-02-01 20:23:48.169: INFO @metrics_hook: train gender_accuracy: 0.942 +-0.001 (110372) 2020-02-01 20:23:50.970: INFO @metrics_hook: valid age_mae: 6.387 +-0.091 (17639) 2020-02-01 20:23:50.971: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 20:23:52.660: INFO @decay_lr : LR updated to `6.401094e-05` 2020-02-01 20:23:52.662: INFO @log_profile : T train: 130.212839 2020-02-01 20:23:52.662: INFO @log_profile : T valid: 6.356238 2020-02-01 20:23:52.662: INFO @log_profile : T read data: 2.837294 2020-02-01 20:23:52.662: INFO @log_profile : T hooks: 5.094765 2020-02-01 20:23:52.662: INFO @main_loop : Epoch 89 done 2020-02-01 20:23:52.662: INFO @main_loop : Training epoch 90 2020-02-01 20:26:13.848: INFO @log_variables: train loss nanmean: 0.748925 2020-02-01 20:26:13.849: INFO @log_variables: train age_loss mean: 5.457072 2020-02-01 20:26:13.849: INFO @log_variables: train gender_loss mean: 0.140838 2020-02-01 20:26:13.849: INFO @log_variables: train age_mae mean: 5.933386 2020-02-01 20:26:13.849: INFO @log_variables: train gender_accuracy mean: 0.943391 2020-02-01 20:26:13.849: INFO @log_variables: train gender_confidence/loss nanmean: 0.055727 2020-02-01 20:26:13.849: INFO @log_variables: train gender_confidence/accuracy mean: 0.842560 2020-02-01 20:26:13.849: INFO @log_variables: train age_confidence/loss mean: 0.069066 2020-02-01 20:26:13.849: INFO @log_variables: train age_confidence/accuracy mean: 0.605552 2020-02-01 20:26:13.849: INFO @log_variables: valid loss nanmean: 0.836283 2020-02-01 20:26:13.849: INFO @log_variables: valid age_loss mean: 5.851113 2020-02-01 20:26:13.849: INFO @log_variables: valid gender_loss mean: 0.199069 2020-02-01 20:26:13.849: INFO @log_variables: valid age_mae mean: 6.332006 2020-02-01 20:26:13.849: INFO @log_variables: valid gender_accuracy mean: 0.920744 2020-02-01 20:26:13.849: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053957 2020-02-01 20:26:13.849: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871931 2020-02-01 20:26:13.849: INFO @log_variables: valid age_confidence/loss mean: 0.069436 2020-02-01 20:26:13.849: INFO @log_variables: valid age_confidence/accuracy mean: 0.564488 2020-02-01 20:26:13.849: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:26:13.857: INFO @metrics_hook: train age_mae: 5.933 +-0.034 (110372) 2020-02-01 20:26:13.864: INFO @metrics_hook: train gender_accuracy: 0.943 +-0.001 (110372) 2020-02-01 20:26:16.689: INFO @metrics_hook: valid age_mae: 6.332 +-0.089 (17639) 2020-02-01 20:26:16.691: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 20:26:18.202: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:26:18.202: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:26:18.202: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.20 2020-02-01 20:26:18.202: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:26:18.335: INFO @evaluate_confidence: Previous accuracy would be: 94.34 2020-02-01 20:26:18.335: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 20:26:18.397: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.5, 97.58, 97.67, 97.75, 97.82, 97.88, 97.96, 98.04, 98.11, 98.18, 98.26, 98.32, 98.39, 98.43, 98.49, 98.55, 98.58, 98.64, 98.68, 98.73, 98.76, 98.79, 98.83, 98.87, 98.9, 98.94, 98.99, 99.04, 99.07, 99.11, 99.14, 99.17, 99.2, 99.24, 99.28, 99.32, 99.35, 99.36, 99.38, 99.41, 99.45, 99.48, 99.51, 99.55] 2020-02-01 20:26:18.398: INFO @evaluate_confidence: Dropped ratios are: [12.9, 13.38, 13.89, 14.39, 14.87, 15.34, 15.82, 16.32, 16.83, 17.33, 17.82, 18.29, 18.78, 19.25, 19.76, 20.27, 20.73, 21.23, 21.74, 22.26, 22.8, 23.3, 23.83, 24.35, 24.85, 25.35, 25.92, 26.48, 27.05, 27.64, 28.23, 28.82, 29.41, 30.03, 30.65, 31.36, 32.1, 32.81, 33.52, 34.26, 35.01, 35.75, 36.54, 37.35] 2020-02-01 20:26:18.447: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:26:18.447: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:26:18.448: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:26:18.448: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:26:18.589: INFO @evaluate_confidence: Previous accuracy would be: 55.69 2020-02-01 20:26:18.589: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:26:18.605: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.61, 65.28, 65.89, 66.56, 67.37, 68.11, 68.94] 2020-02-01 20:26:18.605: INFO @evaluate_confidence: Dropped ratios are: [43.73, 46.82, 49.96, 52.94, 55.93, 58.72, 61.39] 2020-02-01 20:26:18.612: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:26:18.613: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.22 2020-02-01 20:26:18.613: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.22 2020-02-01 20:26:18.613: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.24 2020-02-01 20:26:18.720: INFO @evaluate_confidence: Previous accuracy would be: 92.07 2020-02-01 20:26:18.721: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:26:18.729: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.29, 96.39, 96.51, 96.57, 96.64, 96.73, 96.79, 96.94, 97.09, 97.17, 97.23, 97.32, 97.41, 97.52, 97.64, 97.72, 97.79, 97.9, 97.94, 97.99, 98.05, 98.09, 98.11, 98.15, 98.26, 98.3, 98.37, 98.41, 98.47, 98.5, 98.59, 98.64, 98.67, 98.72, 98.79, 98.86, 98.89, 98.95, 98.98, 99.01] 2020-02-01 20:26:18.729: INFO @evaluate_confidence: Dropped ratios are: [12.78, 13.16, 13.66, 14.14, 14.52, 14.93, 15.29, 15.91, 16.44, 16.82, 17.25, 17.72, 18.07, 18.55, 19.07, 19.55, 20.07, 20.62, 21.12, 21.7, 22.2, 22.78, 23.32, 23.95, 24.51, 25.01, 25.51, 25.99, 26.63, 27.22, 27.84, 28.52, 29.17, 29.81, 30.43, 31.16, 32.0, 32.75, 33.49, 34.41] 2020-02-01 20:26:18.737: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:26:18.738: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 20:26:18.738: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.09 2020-02-01 20:26:18.738: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 20:26:18.869: INFO @evaluate_confidence: Previous accuracy would be: 52.36 2020-02-01 20:26:18.869: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5] 2020-02-01 20:26:18.871: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.86, 59.4, 59.77, 60.24, 61.22] 2020-02-01 20:26:18.871: INFO @evaluate_confidence: Dropped ratios are: [43.57, 47.97, 52.23, 56.4, 60.77] 2020-02-01 20:26:18.932: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:26:19.647: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:26:19.732: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:26:20.210: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:26:20.284: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:26:20.995: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:26:21.084: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:26:21.086: INFO @evaluate_gender-age_model: groups 0 3.826909 1 4.468422 2 5.473365 3 5.808261 4 6.609419 5 6.555803 6 6.687615 7 7.616396 Name: errors, dtype: float64 2020-02-01 20:26:21.087: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:26:21.563: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:26:21.625: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:26:21.626: INFO @evaluate_gender-age_model: groups 0 6.784650 1 5.675233 2 5.923623 3 5.865295 4 7.140357 5 5.157510 6 7.941329 7 12.056915 Name: errors, dtype: float64 2020-02-01 20:26:21.805: INFO @decay_lr : LR updated to `6.369089e-05` 2020-02-01 20:26:21.806: INFO @log_profile : T train: 130.199433 2020-02-01 20:26:21.806: INFO @log_profile : T valid: 5.521610 2020-02-01 20:26:21.806: INFO @log_profile : T read data: 2.855517 2020-02-01 20:26:21.806: INFO @log_profile : T hooks: 10.492563 2020-02-01 20:26:21.806: INFO @main_loop : Epoch 90 done 2020-02-01 20:26:21.806: INFO @main_loop : Training epoch 91 2020-02-01 20:28:33.564: INFO @log_variables: train loss nanmean: 0.743746 2020-02-01 20:28:33.564: INFO @log_variables: train age_loss mean: 5.401508 2020-02-01 20:28:33.564: INFO @log_variables: train gender_loss mean: 0.141035 2020-02-01 20:28:33.564: INFO @log_variables: train age_mae mean: 5.878235 2020-02-01 20:28:33.564: INFO @log_variables: train gender_accuracy mean: 0.943151 2020-02-01 20:28:33.564: INFO @log_variables: train gender_confidence/loss nanmean: 0.055289 2020-02-01 20:28:33.564: INFO @log_variables: train gender_confidence/accuracy mean: 0.841833 2020-02-01 20:28:33.564: INFO @log_variables: train age_confidence/loss mean: 0.069197 2020-02-01 20:28:33.564: INFO @log_variables: train age_confidence/accuracy mean: 0.608037 2020-02-01 20:28:33.564: INFO @log_variables: valid loss nanmean: 0.860769 2020-02-01 20:28:33.564: INFO @log_variables: valid age_loss mean: 5.937546 2020-02-01 20:28:33.564: INFO @log_variables: valid gender_loss mean: 0.217592 2020-02-01 20:28:33.564: INFO @log_variables: valid age_mae mean: 6.417819 2020-02-01 20:28:33.564: INFO @log_variables: valid gender_accuracy mean: 0.912977 2020-02-01 20:28:33.564: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053461 2020-02-01 20:28:33.565: INFO @log_variables: valid gender_confidence/accuracy mean: 0.860310 2020-02-01 20:28:33.565: INFO @log_variables: valid age_confidence/loss mean: 0.069720 2020-02-01 20:28:33.565: INFO @log_variables: valid age_confidence/accuracy mean: 0.562050 2020-02-01 20:28:33.565: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:28:33.572: INFO @metrics_hook: train age_mae: 5.878 +-0.034 (110592) 2020-02-01 20:28:33.579: INFO @metrics_hook: train gender_accuracy: 0.943 +-0.001 (110592) 2020-02-01 20:28:36.291: INFO @metrics_hook: valid age_mae: 6.418 +-0.092 (17639) 2020-02-01 20:28:36.292: INFO @metrics_hook: valid gender_accuracy: 0.913 +-0.004 (17639) 2020-02-01 20:28:37.944: INFO @decay_lr : LR updated to `6.3372434e-05` 2020-02-01 20:28:37.945: INFO @log_profile : T train: 123.673524 2020-02-01 20:28:37.945: INFO @log_profile : T valid: 5.510738 2020-02-01 20:28:37.945: INFO @log_profile : T read data: 1.866899 2020-02-01 20:28:37.945: INFO @log_profile : T hooks: 5.009717 2020-02-01 20:28:37.946: INFO @main_loop : Epoch 91 done 2020-02-01 20:28:37.946: INFO @main_loop : Training epoch 92 2020-02-01 20:30:57.433: INFO @log_variables: train loss nanmean: 0.746107 2020-02-01 20:30:57.433: INFO @log_variables: train age_loss mean: 5.407942 2020-02-01 20:30:57.433: INFO @log_variables: train gender_loss mean: 0.142288 2020-02-01 20:30:57.433: INFO @log_variables: train age_mae mean: 5.884874 2020-02-01 20:30:57.433: INFO @log_variables: train gender_accuracy mean: 0.943364 2020-02-01 20:30:57.433: INFO @log_variables: train gender_confidence/loss nanmean: 0.056012 2020-02-01 20:30:57.433: INFO @log_variables: train gender_confidence/accuracy mean: 0.840983 2020-02-01 20:30:57.433: INFO @log_variables: train age_confidence/loss mean: 0.069111 2020-02-01 20:30:57.433: INFO @log_variables: train age_confidence/accuracy mean: 0.604411 2020-02-01 20:30:57.433: INFO @log_variables: valid loss nanmean: 0.843625 2020-02-01 20:30:57.433: INFO @log_variables: valid age_loss mean: 5.855480 2020-02-01 20:30:57.433: INFO @log_variables: valid gender_loss mean: 0.207124 2020-02-01 20:30:57.433: INFO @log_variables: valid age_mae mean: 6.336021 2020-02-01 20:30:57.434: INFO @log_variables: valid gender_accuracy mean: 0.917852 2020-02-01 20:30:57.434: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053347 2020-02-01 20:30:57.434: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870231 2020-02-01 20:30:57.434: INFO @log_variables: valid age_confidence/loss mean: 0.069667 2020-02-01 20:30:57.434: INFO @log_variables: valid age_confidence/accuracy mean: 0.565452 2020-02-01 20:30:57.434: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:30:57.441: INFO @metrics_hook: train age_mae: 5.885 +-0.034 (110372) 2020-02-01 20:30:57.448: INFO @metrics_hook: train gender_accuracy: 0.943 +-0.001 (110372) 2020-02-01 20:31:00.198: INFO @metrics_hook: valid age_mae: 6.336 +-0.088 (17639) 2020-02-01 20:31:00.199: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 20:31:01.677: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:31:01.677: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:31:01.678: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.21 2020-02-01 20:31:01.678: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:31:01.809: INFO @evaluate_confidence: Previous accuracy would be: 94.34 2020-02-01 20:31:01.810: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 20:31:01.871: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.5, 97.59, 97.67, 97.76, 97.83, 97.89, 97.98, 98.04, 98.12, 98.19, 98.26, 98.32, 98.36, 98.42, 98.49, 98.55, 98.6, 98.65, 98.7, 98.75, 98.79, 98.84, 98.89, 98.92, 98.96, 98.99, 99.03, 99.08, 99.1, 99.13, 99.16, 99.19, 99.22, 99.25, 99.28, 99.31, 99.33, 99.35, 99.38, 99.4, 99.42, 99.45, 99.48, 99.51] 2020-02-01 20:31:01.871: INFO @evaluate_confidence: Dropped ratios are: [12.98, 13.48, 13.96, 14.47, 14.98, 15.49, 15.99, 16.48, 16.98, 17.47, 17.97, 18.45, 18.91, 19.4, 19.88, 20.39, 20.87, 21.39, 21.87, 22.38, 22.87, 23.39, 23.9, 24.4, 24.9, 25.44, 25.98, 26.54, 27.06, 27.59, 28.16, 28.74, 29.37, 29.99, 30.65, 31.31, 31.99, 32.65, 33.36, 34.11, 34.87, 35.62, 36.42, 37.26] 2020-02-01 20:31:01.919: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:31:01.919: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:31:01.919: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:31:01.920: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:31:02.057: INFO @evaluate_confidence: Previous accuracy would be: 55.92 2020-02-01 20:31:02.058: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:31:02.072: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.69, 65.35, 66.17, 66.89, 67.74, 68.59, 69.36] 2020-02-01 20:31:02.073: INFO @evaluate_confidence: Dropped ratios are: [43.56, 46.71, 49.86, 52.79, 55.73, 58.51, 61.06] 2020-02-01 20:31:02.080: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:31:02.080: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 20:31:02.080: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.22 2020-02-01 20:31:02.080: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 20:31:02.185: INFO @evaluate_confidence: Previous accuracy would be: 91.79 2020-02-01 20:31:02.185: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:31:02.194: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.99, 96.08, 96.2, 96.3, 96.39, 96.47, 96.54, 96.63, 96.76, 96.82, 96.93, 97.03, 97.09, 97.2, 97.26, 97.3, 97.36, 97.45, 97.57, 97.63, 97.74, 97.84, 97.87, 97.94, 98.05, 98.12, 98.21, 98.29, 98.37, 98.42, 98.48, 98.54, 98.59, 98.67, 98.74, 98.8, 98.85, 98.89, 98.98, 99.03] 2020-02-01 20:31:02.195: INFO @evaluate_confidence: Dropped ratios are: [13.37, 13.81, 14.27, 14.66, 15.06, 15.42, 15.98, 16.49, 16.95, 17.41, 17.91, 18.54, 19.01, 19.56, 20.04, 20.44, 20.89, 21.37, 21.9, 22.33, 22.95, 23.52, 24.01, 24.58, 25.1, 25.6, 26.27, 26.9, 27.52, 28.15, 28.75, 29.32, 30.0, 30.8, 31.54, 32.3, 33.13, 33.86, 34.67, 35.57] 2020-02-01 20:31:02.202: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:31:02.202: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:31:02.202: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 20:31:02.203: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 20:31:02.332: INFO @evaluate_confidence: Previous accuracy would be: 52.36 2020-02-01 20:31:02.332: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:31:02.333: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.64, 59.02, 59.52, 59.58] 2020-02-01 20:31:02.333: INFO @evaluate_confidence: Dropped ratios are: [44.42, 48.77, 53.2, 57.55] 2020-02-01 20:31:02.388: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:31:03.107: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:31:03.191: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:31:03.647: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:31:03.892: INFO @decay_lr : LR updated to `6.305557e-05` 2020-02-01 20:31:03.894: INFO @log_profile : T train: 129.732865 2020-02-01 20:31:03.894: INFO @log_profile : T valid: 6.286506 2020-02-01 20:31:03.894: INFO @log_profile : T read data: 2.779698 2020-02-01 20:31:03.894: INFO @log_profile : T hooks: 7.073320 2020-02-01 20:31:03.894: INFO @main_loop : Epoch 92 done 2020-02-01 20:31:03.894: INFO @main_loop : Training epoch 93 2020-02-01 20:33:19.690: INFO @log_variables: train loss nanmean: 0.740735 2020-02-01 20:33:19.690: INFO @log_variables: train age_loss mean: 5.361583 2020-02-01 20:33:19.690: INFO @log_variables: train gender_loss mean: 0.140756 2020-02-01 20:33:19.690: INFO @log_variables: train age_mae mean: 5.838480 2020-02-01 20:33:19.690: INFO @log_variables: train gender_accuracy mean: 0.943953 2020-02-01 20:33:19.690: INFO @log_variables: train gender_confidence/loss nanmean: 0.056131 2020-02-01 20:33:19.690: INFO @log_variables: train gender_confidence/accuracy mean: 0.839769 2020-02-01 20:33:19.690: INFO @log_variables: train age_confidence/loss mean: 0.069227 2020-02-01 20:33:19.690: INFO @log_variables: train age_confidence/accuracy mean: 0.610617 2020-02-01 20:33:19.690: INFO @log_variables: valid loss nanmean: 0.846836 2020-02-01 20:33:19.690: INFO @log_variables: valid age_loss mean: 5.837010 2020-02-01 20:33:19.690: INFO @log_variables: valid gender_loss mean: 0.212059 2020-02-01 20:33:19.690: INFO @log_variables: valid age_mae mean: 6.317130 2020-02-01 20:33:19.690: INFO @log_variables: valid gender_accuracy mean: 0.913260 2020-02-01 20:33:19.690: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053590 2020-02-01 20:33:19.690: INFO @log_variables: valid gender_confidence/accuracy mean: 0.861840 2020-02-01 20:33:19.691: INFO @log_variables: valid age_confidence/loss mean: 0.069828 2020-02-01 20:33:19.691: INFO @log_variables: valid age_confidence/accuracy mean: 0.553433 2020-02-01 20:33:19.691: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:33:19.698: INFO @metrics_hook: train age_mae: 5.838 +-0.034 (110372) 2020-02-01 20:33:19.705: INFO @metrics_hook: train gender_accuracy: 0.944 +-0.001 (110372) 2020-02-01 20:33:22.412: INFO @metrics_hook: valid age_mae: 6.317 +-0.089 (17639) 2020-02-01 20:33:22.413: INFO @metrics_hook: valid gender_accuracy: 0.913 +-0.004 (17639) 2020-02-01 20:33:24.030: INFO @decay_lr : LR updated to `6.27403e-05` 2020-02-01 20:33:24.031: INFO @log_profile : T train: 126.808794 2020-02-01 20:33:24.031: INFO @log_profile : T valid: 5.436643 2020-02-01 20:33:24.031: INFO @log_profile : T read data: 2.872317 2020-02-01 20:33:24.031: INFO @log_profile : T hooks: 4.943600 2020-02-01 20:33:24.031: INFO @main_loop : Epoch 93 done 2020-02-01 20:33:24.031: INFO @main_loop : Training epoch 94 2020-02-01 20:35:33.910: INFO @log_variables: train loss nanmean: 0.741925 2020-02-01 20:35:33.910: INFO @log_variables: train age_loss mean: 5.362215 2020-02-01 20:35:33.910: INFO @log_variables: train gender_loss mean: 0.141694 2020-02-01 20:35:33.910: INFO @log_variables: train age_mae mean: 5.838309 2020-02-01 20:35:33.910: INFO @log_variables: train gender_accuracy mean: 0.943811 2020-02-01 20:35:33.910: INFO @log_variables: train gender_confidence/loss nanmean: 0.056249 2020-02-01 20:35:33.910: INFO @log_variables: train gender_confidence/accuracy mean: 0.839536 2020-02-01 20:35:33.910: INFO @log_variables: train age_confidence/loss mean: 0.069389 2020-02-01 20:35:33.910: INFO @log_variables: train age_confidence/accuracy mean: 0.608154 2020-02-01 20:35:33.910: INFO @log_variables: valid loss nanmean: 0.831848 2020-02-01 20:35:33.910: INFO @log_variables: valid age_loss mean: 5.808814 2020-02-01 20:35:33.910: INFO @log_variables: valid gender_loss mean: 0.198622 2020-02-01 20:35:33.910: INFO @log_variables: valid age_mae mean: 6.288769 2020-02-01 20:35:33.910: INFO @log_variables: valid gender_accuracy mean: 0.919043 2020-02-01 20:35:33.911: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053246 2020-02-01 20:35:33.911: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864051 2020-02-01 20:35:33.911: INFO @log_variables: valid age_confidence/loss mean: 0.069962 2020-02-01 20:35:33.911: INFO @log_variables: valid age_confidence/accuracy mean: 0.564261 2020-02-01 20:35:33.911: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:35:33.918: INFO @metrics_hook: train age_mae: 5.838 +-0.033 (110592) 2020-02-01 20:35:33.925: INFO @metrics_hook: train gender_accuracy: 0.944 +-0.001 (110592) 2020-02-01 20:35:36.649: INFO @metrics_hook: valid age_mae: 6.289 +-0.089 (17639) 2020-02-01 20:35:36.650: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 20:35:38.123: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:35:38.123: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:35:38.123: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.38 +- 0.21 2020-02-01 20:35:38.123: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:35:38.250: INFO @evaluate_confidence: Previous accuracy would be: 94.38 2020-02-01 20:35:38.251: INFO @evaluate_confidence: Possible optimal thresholds are: [0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 20:35:38.312: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.49, 97.58, 97.64, 97.73, 97.8, 97.89, 97.96, 98.02, 98.08, 98.15, 98.22, 98.27, 98.33, 98.38, 98.43, 98.49, 98.55, 98.6, 98.63, 98.69, 98.75, 98.79, 98.84, 98.88, 98.92, 98.96, 98.99, 99.02, 99.05, 99.08, 99.12, 99.15, 99.18, 99.22, 99.26, 99.3, 99.33, 99.36, 99.38, 99.41, 99.44, 99.47, 99.5, 99.52] 2020-02-01 20:35:38.312: INFO @evaluate_confidence: Dropped ratios are: [12.96, 13.44, 13.96, 14.44, 14.94, 15.47, 15.99, 16.45, 16.94, 17.41, 17.92, 18.44, 18.96, 19.44, 19.93, 20.42, 20.93, 21.43, 21.94, 22.47, 22.97, 23.49, 24.01, 24.55, 25.09, 25.65, 26.2, 26.77, 27.32, 27.86, 28.46, 29.03, 29.64, 30.31, 30.95, 31.57, 32.23, 32.9, 33.61, 34.31, 35.06, 35.81, 36.6, 37.43] 2020-02-01 20:35:38.361: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:35:38.361: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:35:38.361: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:35:38.362: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:35:38.498: INFO @evaluate_confidence: Previous accuracy would be: 56.21 2020-02-01 20:35:38.498: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:35:38.513: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.02, 65.78, 66.43, 67.07, 67.95, 68.77, 69.54] 2020-02-01 20:35:38.513: INFO @evaluate_confidence: Dropped ratios are: [42.92, 46.06, 49.23, 52.28, 55.2, 57.97, 60.61] 2020-02-01 20:35:38.521: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:35:38.521: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.23 2020-02-01 20:35:38.521: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.22 2020-02-01 20:35:38.521: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.25 2020-02-01 20:35:38.624: INFO @evaluate_confidence: Previous accuracy would be: 91.90 2020-02-01 20:35:38.625: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:35:38.633: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.21, 96.32, 96.45, 96.6, 96.68, 96.8, 96.93, 97.06, 97.16, 97.26, 97.33, 97.38, 97.4, 97.42, 97.5, 97.59, 97.7, 97.78, 97.87, 97.95, 97.99, 98.05, 98.13, 98.16, 98.2, 98.23, 98.29, 98.33, 98.4, 98.45, 98.5, 98.55, 98.6, 98.7, 98.71, 98.76, 98.81, 98.84, 98.91, 98.94, 99.01] 2020-02-01 20:35:38.633: INFO @evaluate_confidence: Dropped ratios are: [13.52, 14.01, 14.45, 14.87, 15.29, 15.75, 16.29, 16.79, 17.25, 17.75, 18.13, 18.56, 19.04, 19.47, 19.92, 20.47, 20.93, 21.42, 21.92, 22.47, 22.9, 23.43, 23.99, 24.55, 25.0, 25.64, 26.2, 26.78, 27.21, 27.81, 28.36, 28.93, 29.43, 30.17, 30.81, 31.5, 32.08, 32.83, 33.62, 34.42, 35.26] 2020-02-01 20:35:38.641: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:35:38.641: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:35:38.641: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 20:35:38.641: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 20:35:38.770: INFO @evaluate_confidence: Previous accuracy would be: 52.62 2020-02-01 20:35:38.770: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:35:38.772: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.98, 59.29, 59.45, 59.93] 2020-02-01 20:35:38.772: INFO @evaluate_confidence: Dropped ratios are: [43.32, 47.83, 52.13, 56.51] 2020-02-01 20:35:38.825: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:35:39.560: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:35:39.645: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:35:40.117: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:35:40.363: INFO @decay_lr : LR updated to `6.2426596e-05` 2020-02-01 20:35:40.364: INFO @log_profile : T train: 121.777666 2020-02-01 20:35:40.364: INFO @log_profile : T valid: 5.441731 2020-02-01 20:35:40.364: INFO @log_profile : T read data: 1.956691 2020-02-01 20:35:40.364: INFO @log_profile : T hooks: 7.079089 2020-02-01 20:35:40.364: INFO @main_loop : Epoch 94 done 2020-02-01 20:35:40.364: INFO @main_loop : Training epoch 95 2020-02-01 20:37:51.100: INFO @log_variables: train loss nanmean: 0.739825 2020-02-01 20:37:51.100: INFO @log_variables: train age_loss mean: 5.372900 2020-02-01 20:37:51.100: INFO @log_variables: train gender_loss mean: 0.139240 2020-02-01 20:37:51.101: INFO @log_variables: train age_mae mean: 5.849595 2020-02-01 20:37:51.101: INFO @log_variables: train gender_accuracy mean: 0.944279 2020-02-01 20:37:51.101: INFO @log_variables: train gender_confidence/loss nanmean: 0.055612 2020-02-01 20:37:51.101: INFO @log_variables: train gender_confidence/accuracy mean: 0.843058 2020-02-01 20:37:51.101: INFO @log_variables: train age_confidence/loss mean: 0.069186 2020-02-01 20:37:51.101: INFO @log_variables: train age_confidence/accuracy mean: 0.609539 2020-02-01 20:37:51.101: INFO @log_variables: valid loss nanmean: 0.851224 2020-02-01 20:37:51.101: INFO @log_variables: valid age_loss mean: 5.870992 2020-02-01 20:37:51.101: INFO @log_variables: valid gender_loss mean: 0.211414 2020-02-01 20:37:51.101: INFO @log_variables: valid age_mae mean: 6.351993 2020-02-01 20:37:51.101: INFO @log_variables: valid gender_accuracy mean: 0.918873 2020-02-01 20:37:51.101: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055558 2020-02-01 20:37:51.101: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875503 2020-02-01 20:37:51.101: INFO @log_variables: valid age_confidence/loss mean: 0.069745 2020-02-01 20:37:51.101: INFO @log_variables: valid age_confidence/accuracy mean: 0.560859 2020-02-01 20:37:51.101: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:37:51.108: INFO @metrics_hook: train age_mae: 5.850 +-0.034 (110372) 2020-02-01 20:37:51.115: INFO @metrics_hook: train gender_accuracy: 0.944 +-0.001 (110372) 2020-02-01 20:37:53.844: INFO @metrics_hook: valid age_mae: 6.352 +-0.089 (17639) 2020-02-01 20:37:53.845: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 20:37:55.476: INFO @decay_lr : LR updated to `6.2114465e-05` 2020-02-01 20:37:55.477: INFO @log_profile : T train: 121.667229 2020-02-01 20:37:55.477: INFO @log_profile : T valid: 5.513938 2020-02-01 20:37:55.477: INFO @log_profile : T read data: 2.886379 2020-02-01 20:37:55.477: INFO @log_profile : T hooks: 4.969570 2020-02-01 20:37:55.477: INFO @main_loop : Epoch 95 done 2020-02-01 20:37:55.477: INFO @main_loop : Training epoch 96 2020-02-01 20:40:06.076: INFO @log_variables: train loss nanmean: 0.738914 2020-02-01 20:40:06.076: INFO @log_variables: train age_loss mean: 5.362738 2020-02-01 20:40:06.076: INFO @log_variables: train gender_loss mean: 0.138859 2020-02-01 20:40:06.076: INFO @log_variables: train age_mae mean: 5.838810 2020-02-01 20:40:06.076: INFO @log_variables: train gender_accuracy mean: 0.945330 2020-02-01 20:40:06.076: INFO @log_variables: train gender_confidence/loss nanmean: 0.055709 2020-02-01 20:40:06.076: INFO @log_variables: train gender_confidence/accuracy mean: 0.842261 2020-02-01 20:40:06.076: INFO @log_variables: train age_confidence/loss mean: 0.069448 2020-02-01 20:40:06.076: INFO @log_variables: train age_confidence/accuracy mean: 0.604782 2020-02-01 20:40:06.076: INFO @log_variables: valid loss nanmean: 0.828171 2020-02-01 20:40:06.076: INFO @log_variables: valid age_loss mean: 5.875560 2020-02-01 20:40:06.076: INFO @log_variables: valid gender_loss mean: 0.190689 2020-02-01 20:40:06.076: INFO @log_variables: valid age_mae mean: 6.356285 2020-02-01 20:40:06.076: INFO @log_variables: valid gender_accuracy mean: 0.921878 2020-02-01 20:40:06.076: INFO @log_variables: valid gender_confidence/loss nanmean: 0.050633 2020-02-01 20:40:06.076: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868360 2020-02-01 20:40:06.077: INFO @log_variables: valid age_confidence/loss mean: 0.070042 2020-02-01 20:40:06.077: INFO @log_variables: valid age_confidence/accuracy mean: 0.546119 2020-02-01 20:40:06.077: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:40:06.083: INFO @metrics_hook: train age_mae: 5.839 +-0.034 (110372) 2020-02-01 20:40:06.090: INFO @metrics_hook: train gender_accuracy: 0.945 +-0.001 (110372) 2020-02-01 20:40:08.830: INFO @metrics_hook: valid age_mae: 6.356 +-0.091 (17639) 2020-02-01 20:40:08.831: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 20:40:10.285: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:40:10.285: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:40:10.285: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 20:40:10.285: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:40:10.413: INFO @evaluate_confidence: Previous accuracy would be: 94.53 2020-02-01 20:40:10.414: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:40:10.473: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.51, 97.59, 97.68, 97.76, 97.85, 97.92, 97.99, 98.06, 98.11, 98.17, 98.23, 98.28, 98.34, 98.39, 98.45, 98.5, 98.56, 98.6, 98.66, 98.71, 98.76, 98.8, 98.83, 98.89, 98.92, 98.97, 99.02, 99.06, 99.1, 99.14, 99.18, 99.21, 99.24, 99.28, 99.31, 99.33, 99.36, 99.38, 99.4, 99.43, 99.45, 99.48, 99.5, 99.52] 2020-02-01 20:40:10.473: INFO @evaluate_confidence: Dropped ratios are: [12.4, 12.88, 13.4, 13.89, 14.37, 14.87, 15.36, 15.82, 16.3, 16.78, 17.26, 17.68, 18.14, 18.62, 19.1, 19.6, 20.06, 20.56, 21.01, 21.49, 22.02, 22.51, 22.99, 23.54, 24.04, 24.58, 25.13, 25.65, 26.17, 26.72, 27.27, 27.84, 28.43, 29.02, 29.64, 30.25, 30.89, 31.55, 32.25, 32.93, 33.64, 34.38, 35.19, 36.02] 2020-02-01 20:40:10.522: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:40:10.522: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:40:10.522: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:40:10.522: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:40:10.658: INFO @evaluate_confidence: Previous accuracy would be: 56.33 2020-02-01 20:40:10.658: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:40:10.673: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [64.83, 65.6, 66.35, 67.11, 67.91, 68.67, 69.62] 2020-02-01 20:40:10.673: INFO @evaluate_confidence: Dropped ratios are: [42.97, 46.12, 49.2, 52.08, 54.98, 57.85, 60.53] 2020-02-01 20:40:10.680: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:40:10.681: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 20:40:10.681: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.22 2020-02-01 20:40:10.681: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 20:40:10.784: INFO @evaluate_confidence: Previous accuracy would be: 92.19 2020-02-01 20:40:10.784: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:40:10.793: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.38, 96.48, 96.56, 96.64, 96.71, 96.86, 96.96, 97.03, 97.13, 97.2, 97.28, 97.34, 97.4, 97.45, 97.56, 97.6, 97.66, 97.71, 97.82, 97.88, 98.0, 98.07, 98.14, 98.2, 98.25, 98.31, 98.38, 98.44, 98.53, 98.59, 98.71, 98.76, 98.82, 98.88, 98.91, 98.94, 98.97, 99.02, 99.06, 99.08, 99.14, 99.22, 99.28] 2020-02-01 20:40:10.793: INFO @evaluate_confidence: Dropped ratios are: [13.03, 13.46, 13.84, 14.22, 14.6, 15.08, 15.52, 15.99, 16.3, 16.75, 17.13, 17.5, 17.95, 18.33, 18.79, 19.15, 19.5, 20.01, 20.51, 20.94, 21.53, 21.97, 22.41, 22.94, 23.4, 23.84, 24.24, 24.75, 25.41, 25.89, 26.4, 27.08, 27.72, 28.34, 28.99, 29.58, 30.22, 30.86, 31.58, 32.34, 33.18, 34.02, 35.01] 2020-02-01 20:40:10.801: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:40:10.801: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 20:40:10.801: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 20:40:10.801: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 20:40:10.931: INFO @evaluate_confidence: Previous accuracy would be: 52.49 2020-02-01 20:40:10.931: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 20:40:10.933: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.08, 58.38, 58.36, 58.65] 2020-02-01 20:40:10.933: INFO @evaluate_confidence: Dropped ratios are: [45.67, 50.18, 54.88, 58.97] 2020-02-01 20:40:10.983: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:40:11.672: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:40:11.756: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:40:12.205: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:40:12.451: INFO @decay_lr : LR updated to `6.180389e-05` 2020-02-01 20:40:12.452: INFO @log_profile : T train: 121.557027 2020-02-01 20:40:12.452: INFO @log_profile : T valid: 5.430991 2020-02-01 20:40:12.452: INFO @log_profile : T read data: 2.939469 2020-02-01 20:40:12.452: INFO @log_profile : T hooks: 6.969537 2020-02-01 20:40:12.452: INFO @main_loop : Epoch 96 done 2020-02-01 20:40:12.452: INFO @main_loop : Training epoch 97 2020-02-01 20:42:30.691: INFO @log_variables: train loss nanmean: 0.733477 2020-02-01 20:42:30.691: INFO @log_variables: train age_loss mean: 5.332253 2020-02-01 20:42:30.691: INFO @log_variables: train gender_loss mean: 0.136377 2020-02-01 20:42:30.691: INFO @log_variables: train age_mae mean: 5.809180 2020-02-01 20:42:30.691: INFO @log_variables: train gender_accuracy mean: 0.946117 2020-02-01 20:42:30.691: INFO @log_variables: train gender_confidence/loss nanmean: 0.055203 2020-02-01 20:42:30.691: INFO @log_variables: train gender_confidence/accuracy mean: 0.845712 2020-02-01 20:42:30.691: INFO @log_variables: train age_confidence/loss mean: 0.069544 2020-02-01 20:42:30.692: INFO @log_variables: train age_confidence/accuracy mean: 0.605939 2020-02-01 20:42:30.692: INFO @log_variables: valid loss nanmean: 0.832292 2020-02-01 20:42:30.692: INFO @log_variables: valid age_loss mean: 5.809305 2020-02-01 20:42:30.692: INFO @log_variables: valid gender_loss mean: 0.198350 2020-02-01 20:42:30.692: INFO @log_variables: valid age_mae mean: 6.288770 2020-02-01 20:42:30.692: INFO @log_variables: valid gender_accuracy mean: 0.920857 2020-02-01 20:42:30.692: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053640 2020-02-01 20:42:30.692: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866092 2020-02-01 20:42:30.692: INFO @log_variables: valid age_confidence/loss mean: 0.070215 2020-02-01 20:42:30.692: INFO @log_variables: valid age_confidence/accuracy mean: 0.557628 2020-02-01 20:42:30.692: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:42:30.699: INFO @metrics_hook: train age_mae: 5.809 +-0.033 (110592) 2020-02-01 20:42:30.706: INFO @metrics_hook: train gender_accuracy: 0.946 +-0.001 (110592) 2020-02-01 20:42:33.438: INFO @metrics_hook: valid age_mae: 6.289 +-0.090 (17639) 2020-02-01 20:42:33.440: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 20:42:35.080: INFO @decay_lr : LR updated to `6.149487e-05` 2020-02-01 20:42:35.081: INFO @log_profile : T train: 130.065173 2020-02-01 20:42:35.081: INFO @log_profile : T valid: 5.643902 2020-02-01 20:42:35.081: INFO @log_profile : T read data: 1.839782 2020-02-01 20:42:35.081: INFO @log_profile : T hooks: 5.003713 2020-02-01 20:42:35.081: INFO @main_loop : Epoch 97 done 2020-02-01 20:42:35.081: INFO @main_loop : Training epoch 98 2020-02-01 20:44:45.841: INFO @log_variables: train loss nanmean: 0.737571 2020-02-01 20:44:45.841: INFO @log_variables: train age_loss mean: 5.342101 2020-02-01 20:44:45.841: INFO @log_variables: train gender_loss mean: 0.139559 2020-02-01 20:44:45.841: INFO @log_variables: train age_mae mean: 5.818242 2020-02-01 20:44:45.841: INFO @log_variables: train gender_accuracy mean: 0.944515 2020-02-01 20:44:45.841: INFO @log_variables: train gender_confidence/loss nanmean: 0.055551 2020-02-01 20:44:45.841: INFO @log_variables: train gender_confidence/accuracy mean: 0.842967 2020-02-01 20:44:45.841: INFO @log_variables: train age_confidence/loss mean: 0.069503 2020-02-01 20:44:45.841: INFO @log_variables: train age_confidence/accuracy mean: 0.608216 2020-02-01 20:44:45.842: INFO @log_variables: valid loss nanmean: 0.834832 2020-02-01 20:44:45.842: INFO @log_variables: valid age_loss mean: 5.875166 2020-02-01 20:44:45.842: INFO @log_variables: valid gender_loss mean: 0.193232 2020-02-01 20:44:45.842: INFO @log_variables: valid age_mae mean: 6.355199 2020-02-01 20:44:45.842: INFO @log_variables: valid gender_accuracy mean: 0.921538 2020-02-01 20:44:45.842: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054596 2020-02-01 20:44:45.842: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871421 2020-02-01 20:44:45.842: INFO @log_variables: valid age_confidence/loss mean: 0.070464 2020-02-01 20:44:45.842: INFO @log_variables: valid age_confidence/accuracy mean: 0.556551 2020-02-01 20:44:45.842: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:44:45.849: INFO @metrics_hook: train age_mae: 5.818 +-0.034 (110372) 2020-02-01 20:44:45.856: INFO @metrics_hook: train gender_accuracy: 0.945 +-0.001 (110372) 2020-02-01 20:44:48.582: INFO @metrics_hook: valid age_mae: 6.355 +-0.091 (17639) 2020-02-01 20:44:48.584: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 20:44:50.060: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:44:50.060: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:44:50.060: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 20:44:50.060: INFO @evaluate_confidence: Average confidence of all samples 0.78 +- 0.26 2020-02-01 20:44:50.190: INFO @evaluate_confidence: Previous accuracy would be: 94.45 2020-02-01 20:44:50.190: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:44:50.250: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.56, 97.66, 97.74, 97.82, 97.88, 97.95, 98.01, 98.08, 98.14, 98.19, 98.25, 98.31, 98.36, 98.43, 98.49, 98.53, 98.57, 98.62, 98.66, 98.7, 98.74, 98.79, 98.84, 98.89, 98.92, 98.95, 98.97, 99.01, 99.06, 99.09, 99.13, 99.17, 99.21, 99.25, 99.29, 99.32, 99.34, 99.38, 99.4, 99.44, 99.45, 99.48, 99.5, 99.53] 2020-02-01 20:44:50.250: INFO @evaluate_confidence: Dropped ratios are: [12.47, 12.99, 13.47, 13.96, 14.44, 14.91, 15.39, 15.88, 16.35, 16.81, 17.28, 17.76, 18.25, 18.71, 19.18, 19.66, 20.17, 20.65, 21.16, 21.64, 22.18, 22.67, 23.16, 23.71, 24.19, 24.75, 25.29, 25.82, 26.35, 26.96, 27.54, 28.11, 28.73, 29.31, 29.92, 30.54, 31.18, 31.91, 32.54, 33.22, 33.93, 34.7, 35.45, 36.25] 2020-02-01 20:44:50.299: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:44:50.299: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:44:50.300: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:44:50.300: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:44:50.437: INFO @evaluate_confidence: Previous accuracy would be: 56.51 2020-02-01 20:44:50.437: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:44:50.452: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.25, 65.92, 66.6, 67.37, 68.02, 68.78, 69.45] 2020-02-01 20:44:50.452: INFO @evaluate_confidence: Dropped ratios are: [42.51, 45.57, 48.6, 51.57, 54.53, 57.37, 60.01] 2020-02-01 20:44:50.460: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:44:50.460: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 20:44:50.460: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.22 2020-02-01 20:44:50.460: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 20:44:50.564: INFO @evaluate_confidence: Previous accuracy would be: 92.15 2020-02-01 20:44:50.564: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:44:50.573: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.22, 96.31, 96.43, 96.54, 96.63, 96.76, 96.86, 96.94, 97.03, 97.13, 97.18, 97.28, 97.35, 97.42, 97.47, 97.55, 97.62, 97.65, 97.73, 97.82, 97.94, 98.0, 98.08, 98.15, 98.19, 98.25, 98.36, 98.45, 98.51, 98.57, 98.61, 98.68, 98.75, 98.77, 98.79, 98.85, 98.9, 98.98, 99.03, 99.1, 99.16, 99.23] 2020-02-01 20:44:50.573: INFO @evaluate_confidence: Dropped ratios are: [12.91, 13.29, 13.67, 14.09, 14.47, 14.96, 15.39, 15.81, 16.15, 16.55, 16.94, 17.34, 17.68, 18.03, 18.34, 18.94, 19.4, 19.85, 20.28, 20.85, 21.44, 21.95, 22.46, 22.97, 23.49, 24.04, 24.59, 25.17, 25.81, 26.28, 26.83, 27.39, 28.06, 28.75, 29.47, 30.26, 30.98, 31.68, 32.55, 33.41, 34.38, 35.28] 2020-02-01 20:44:50.580: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:44:50.581: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 20:44:50.581: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 20:44:50.581: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.11 2020-02-01 20:44:50.709: INFO @evaluate_confidence: Previous accuracy would be: 52.74 2020-02-01 20:44:50.710: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 20:44:50.711: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.2, 58.45, 58.48, 58.54] 2020-02-01 20:44:50.711: INFO @evaluate_confidence: Dropped ratios are: [46.16, 50.35, 54.62, 58.63] 2020-02-01 20:44:50.765: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:44:51.464: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:44:51.549: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:44:52.011: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:44:52.258: INFO @decay_lr : LR updated to `6.11874e-05` 2020-02-01 20:44:52.260: INFO @log_profile : T train: 121.737077 2020-02-01 20:44:52.260: INFO @log_profile : T valid: 5.424753 2020-02-01 20:44:52.260: INFO @log_profile : T read data: 2.892904 2020-02-01 20:44:52.260: INFO @log_profile : T hooks: 7.046588 2020-02-01 20:44:52.260: INFO @main_loop : Epoch 98 done 2020-02-01 20:44:52.260: INFO @main_loop : Training epoch 99 2020-02-01 20:47:03.032: INFO @log_variables: train loss nanmean: 0.732911 2020-02-01 20:47:03.032: INFO @log_variables: train age_loss mean: 5.307240 2020-02-01 20:47:03.032: INFO @log_variables: train gender_loss mean: 0.137445 2020-02-01 20:47:03.033: INFO @log_variables: train age_mae mean: 5.783501 2020-02-01 20:47:03.033: INFO @log_variables: train gender_accuracy mean: 0.945965 2020-02-01 20:47:03.033: INFO @log_variables: train gender_confidence/loss nanmean: 0.055805 2020-02-01 20:47:03.033: INFO @log_variables: train gender_confidence/accuracy mean: 0.842967 2020-02-01 20:47:03.033: INFO @log_variables: train age_confidence/loss mean: 0.069679 2020-02-01 20:47:03.033: INFO @log_variables: train age_confidence/accuracy mean: 0.607047 2020-02-01 20:47:03.033: INFO @log_variables: valid loss nanmean: 0.837911 2020-02-01 20:47:03.033: INFO @log_variables: valid age_loss mean: 5.862042 2020-02-01 20:47:03.033: INFO @log_variables: valid gender_loss mean: 0.197496 2020-02-01 20:47:03.033: INFO @log_variables: valid age_mae mean: 6.341594 2020-02-01 20:47:03.033: INFO @log_variables: valid gender_accuracy mean: 0.920460 2020-02-01 20:47:03.033: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054984 2020-02-01 20:47:03.033: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869380 2020-02-01 20:47:03.033: INFO @log_variables: valid age_confidence/loss mean: 0.070471 2020-02-01 20:47:03.033: INFO @log_variables: valid age_confidence/accuracy mean: 0.564148 2020-02-01 20:47:03.033: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:47:03.040: INFO @metrics_hook: train age_mae: 5.784 +-0.033 (110372) 2020-02-01 20:47:03.048: INFO @metrics_hook: train gender_accuracy: 0.946 +-0.001 (110372) 2020-02-01 20:47:05.752: INFO @metrics_hook: valid age_mae: 6.342 +-0.092 (17639) 2020-02-01 20:47:05.754: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 20:47:07.388: INFO @decay_lr : LR updated to `6.0881463e-05` 2020-02-01 20:47:07.389: INFO @log_profile : T train: 121.810041 2020-02-01 20:47:07.389: INFO @log_profile : T valid: 5.447236 2020-02-01 20:47:07.389: INFO @log_profile : T read data: 2.828696 2020-02-01 20:47:07.389: INFO @log_profile : T hooks: 4.965897 2020-02-01 20:47:07.389: INFO @main_loop : Epoch 99 done 2020-02-01 20:47:07.389: INFO @main_loop : Training epoch 100 2020-02-01 20:49:18.683: INFO @log_variables: train loss nanmean: 0.727550 2020-02-01 20:49:18.683: INFO @log_variables: train age_loss mean: 5.311700 2020-02-01 20:49:18.683: INFO @log_variables: train gender_loss mean: 0.131957 2020-02-01 20:49:18.683: INFO @log_variables: train age_mae mean: 5.788532 2020-02-01 20:49:18.683: INFO @log_variables: train gender_accuracy mean: 0.947528 2020-02-01 20:49:18.683: INFO @log_variables: train gender_confidence/loss nanmean: 0.054972 2020-02-01 20:49:18.683: INFO @log_variables: train gender_confidence/accuracy mean: 0.845993 2020-02-01 20:49:18.683: INFO @log_variables: train age_confidence/loss mean: 0.069735 2020-02-01 20:49:18.683: INFO @log_variables: train age_confidence/accuracy mean: 0.605903 2020-02-01 20:49:18.684: INFO @log_variables: valid loss nanmean: 0.850199 2020-02-01 20:49:18.684: INFO @log_variables: valid age_loss mean: 5.892229 2020-02-01 20:49:18.684: INFO @log_variables: valid gender_loss mean: 0.208467 2020-02-01 20:49:18.684: INFO @log_variables: valid age_mae mean: 6.371935 2020-02-01 20:49:18.684: INFO @log_variables: valid gender_accuracy mean: 0.917002 2020-02-01 20:49:18.684: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055162 2020-02-01 20:49:18.684: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868360 2020-02-01 20:49:18.684: INFO @log_variables: valid age_confidence/loss mean: 0.069864 2020-02-01 20:49:18.684: INFO @log_variables: valid age_confidence/accuracy mean: 0.557118 2020-02-01 20:49:18.684: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:49:18.691: INFO @metrics_hook: train age_mae: 5.789 +-0.033 (110592) 2020-02-01 20:49:18.699: INFO @metrics_hook: train gender_accuracy: 0.948 +-0.001 (110592) 2020-02-01 20:49:21.425: INFO @metrics_hook: valid age_mae: 6.372 +-0.091 (17639) 2020-02-01 20:49:21.426: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 20:49:22.882: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:49:22.882: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:49:22.882: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 20:49:22.882: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 20:49:23.015: INFO @evaluate_confidence: Previous accuracy would be: 94.75 2020-02-01 20:49:23.016: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:49:23.075: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.69, 97.77, 97.85, 97.93, 98.0, 98.08, 98.15, 98.21, 98.27, 98.32, 98.36, 98.42, 98.47, 98.52, 98.57, 98.61, 98.68, 98.73, 98.78, 98.81, 98.86, 98.9, 98.94, 98.97, 99.0, 99.04, 99.08, 99.11, 99.14, 99.18, 99.21, 99.24, 99.27, 99.29, 99.33, 99.36, 99.38, 99.41, 99.43, 99.46, 99.49, 99.5, 99.51, 99.54] 2020-02-01 20:49:23.075: INFO @evaluate_confidence: Dropped ratios are: [12.01, 12.49, 13.02, 13.53, 14.01, 14.52, 15.0, 15.45, 15.92, 16.41, 16.87, 17.31, 17.78, 18.23, 18.71, 19.22, 19.72, 20.19, 20.65, 21.16, 21.65, 22.12, 22.61, 23.1, 23.62, 24.15, 24.67, 25.19, 25.78, 26.34, 26.87, 27.42, 27.99, 28.56, 29.18, 29.83, 30.47, 31.1, 31.79, 32.53, 33.22, 34.0, 34.78, 35.57] 2020-02-01 20:49:23.124: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:49:23.124: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:49:23.124: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:49:23.124: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.13 2020-02-01 20:49:23.259: INFO @evaluate_confidence: Previous accuracy would be: 56.60 2020-02-01 20:49:23.259: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:49:23.273: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.0, 65.69, 66.45, 67.12, 67.83, 68.56, 69.37] 2020-02-01 20:49:23.274: INFO @evaluate_confidence: Dropped ratios are: [42.08, 45.22, 48.23, 51.23, 54.22, 57.11, 59.78] 2020-02-01 20:49:23.281: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:49:23.281: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 20:49:23.282: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.22 2020-02-01 20:49:23.282: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.24 2020-02-01 20:49:23.383: INFO @evaluate_confidence: Previous accuracy would be: 91.70 2020-02-01 20:49:23.383: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 20:49:23.391: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.83, 95.94, 96.09, 96.25, 96.36, 96.44, 96.55, 96.64, 96.75, 96.87, 96.97, 97.04, 97.1, 97.22, 97.35, 97.45, 97.56, 97.64, 97.71, 97.79, 97.82, 97.9, 98.01, 98.06, 98.11, 98.17, 98.22, 98.28, 98.36, 98.43, 98.46, 98.53, 98.54, 98.59, 98.67, 98.73, 98.76, 98.82, 98.89, 98.95] 2020-02-01 20:49:23.391: INFO @evaluate_confidence: Dropped ratios are: [13.49, 13.94, 14.39, 14.88, 15.3, 15.71, 16.13, 16.59, 17.03, 17.54, 17.99, 18.44, 18.84, 19.33, 19.85, 20.35, 20.85, 21.38, 21.85, 22.44, 22.88, 23.43, 23.92, 24.47, 25.09, 25.69, 26.32, 27.01, 27.61, 28.3, 28.92, 29.65, 30.33, 31.05, 31.86, 32.63, 33.49, 34.43, 35.4, 36.39] 2020-02-01 20:49:23.399: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:49:23.399: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.51 +- 0.11 2020-02-01 20:49:23.399: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.10 2020-02-01 20:49:23.399: INFO @evaluate_confidence: Average confidence of all samples 0.49 +- 0.10 2020-02-01 20:49:23.527: INFO @evaluate_confidence: Previous accuracy would be: 52.58 2020-02-01 20:49:23.527: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5] 2020-02-01 20:49:23.528: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.74, 59.17, 59.78, 60.49] 2020-02-01 20:49:23.529: INFO @evaluate_confidence: Dropped ratios are: [46.41, 50.97, 56.01, 60.47] 2020-02-01 20:49:23.579: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:49:24.266: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:49:24.348: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:49:24.792: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:49:24.866: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:49:25.554: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:49:25.635: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:49:25.637: INFO @evaluate_gender-age_model: groups 0 3.823208 1 4.270717 2 5.404849 3 5.764044 4 6.525709 5 6.386398 6 6.403304 7 7.275281 Name: errors, dtype: float64 2020-02-01 20:49:25.638: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:49:26.089: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:49:26.150: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 20:49:26.151: INFO @evaluate_gender-age_model: groups 0 5.891916 1 5.013558 2 5.249733 3 5.484223 4 7.210558 5 5.526215 6 8.983971 7 13.086028 Name: errors, dtype: float64 2020-02-01 20:49:26.324: INFO @decay_lr : LR updated to `6.0577055e-05` 2020-02-01 20:49:26.325: INFO @log_profile : T train: 121.594034 2020-02-01 20:49:26.325: INFO @log_profile : T valid: 5.371870 2020-02-01 20:49:26.325: INFO @log_profile : T read data: 1.926848 2020-02-01 20:49:26.325: INFO @log_profile : T hooks: 9.967729 2020-02-01 20:49:26.325: INFO @main_loop : Epoch 100 done 2020-02-01 20:49:26.325: INFO @main_loop : Training epoch 101 2020-02-01 20:51:37.278: INFO @log_variables: train loss nanmean: 0.729877 2020-02-01 20:51:37.278: INFO @log_variables: train age_loss mean: 5.331548 2020-02-01 20:51:37.278: INFO @log_variables: train gender_loss mean: 0.132660 2020-02-01 20:51:37.279: INFO @log_variables: train age_mae mean: 5.808282 2020-02-01 20:51:37.279: INFO @log_variables: train gender_accuracy mean: 0.947269 2020-02-01 20:51:37.279: INFO @log_variables: train gender_confidence/loss nanmean: 0.055081 2020-02-01 20:51:37.279: INFO @log_variables: train gender_confidence/accuracy mean: 0.844209 2020-02-01 20:51:37.279: INFO @log_variables: train age_confidence/loss mean: 0.069510 2020-02-01 20:51:37.279: INFO @log_variables: train age_confidence/accuracy mean: 0.605905 2020-02-01 20:51:37.279: INFO @log_variables: valid loss nanmean: 0.842802 2020-02-01 20:51:37.279: INFO @log_variables: valid age_loss mean: 5.769516 2020-02-01 20:51:37.279: INFO @log_variables: valid gender_loss mean: 0.211816 2020-02-01 20:51:37.279: INFO @log_variables: valid age_mae mean: 6.250117 2020-02-01 20:51:37.279: INFO @log_variables: valid gender_accuracy mean: 0.919043 2020-02-01 20:51:37.279: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055655 2020-02-01 20:51:37.279: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869664 2020-02-01 20:51:37.279: INFO @log_variables: valid age_confidence/loss mean: 0.070085 2020-02-01 20:51:37.279: INFO @log_variables: valid age_confidence/accuracy mean: 0.571574 2020-02-01 20:51:37.279: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:51:37.286: INFO @metrics_hook: train age_mae: 5.808 +-0.033 (110372) 2020-02-01 20:51:37.294: INFO @metrics_hook: train gender_accuracy: 0.947 +-0.001 (110372) 2020-02-01 20:51:39.994: INFO @metrics_hook: valid age_mae: 6.250 +-0.088 (17639) 2020-02-01 20:51:39.995: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 20:51:41.630: INFO @decay_lr : LR updated to `6.027417e-05` 2020-02-01 20:51:41.632: INFO @log_profile : T train: 121.866556 2020-02-01 20:51:41.632: INFO @log_profile : T valid: 5.512504 2020-02-01 20:51:41.632: INFO @log_profile : T read data: 2.878447 2020-02-01 20:51:41.632: INFO @log_profile : T hooks: 4.971487 2020-02-01 20:51:41.632: INFO @main_loop : Epoch 101 done 2020-02-01 20:51:41.632: INFO @main_loop : Training epoch 102 2020-02-01 20:53:59.299: INFO @log_variables: train loss nanmean: 0.723109 2020-02-01 20:53:59.299: INFO @log_variables: train age_loss mean: 5.277113 2020-02-01 20:53:59.299: INFO @log_variables: train gender_loss mean: 0.130891 2020-02-01 20:53:59.299: INFO @log_variables: train age_mae mean: 5.753307 2020-02-01 20:53:59.299: INFO @log_variables: train gender_accuracy mean: 0.948538 2020-02-01 20:53:59.299: INFO @log_variables: train gender_confidence/loss nanmean: 0.054636 2020-02-01 20:53:59.299: INFO @log_variables: train gender_confidence/accuracy mean: 0.844317 2020-02-01 20:53:59.300: INFO @log_variables: train age_confidence/loss mean: 0.069743 2020-02-01 20:53:59.300: INFO @log_variables: train age_confidence/accuracy mean: 0.607020 2020-02-01 20:53:59.300: INFO @log_variables: valid loss nanmean: 0.849562 2020-02-01 20:53:59.300: INFO @log_variables: valid age_loss mean: 5.925034 2020-02-01 20:53:59.300: INFO @log_variables: valid gender_loss mean: 0.206373 2020-02-01 20:53:59.300: INFO @log_variables: valid age_mae mean: 6.405973 2020-02-01 20:53:59.300: INFO @log_variables: valid gender_accuracy mean: 0.917626 2020-02-01 20:53:59.300: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053679 2020-02-01 20:53:59.300: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862691 2020-02-01 20:53:59.300: INFO @log_variables: valid age_confidence/loss mean: 0.069632 2020-02-01 20:53:59.300: INFO @log_variables: valid age_confidence/accuracy mean: 0.561710 2020-02-01 20:53:59.300: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:53:59.307: INFO @metrics_hook: train age_mae: 5.753 +-0.033 (110372) 2020-02-01 20:53:59.314: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110372) 2020-02-01 20:54:01.999: INFO @metrics_hook: valid age_mae: 6.406 +-0.091 (17639) 2020-02-01 20:54:02.000: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 20:54:03.444: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:54:03.444: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:54:03.444: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.20 2020-02-01 20:54:03.445: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 20:54:03.567: INFO @evaluate_confidence: Previous accuracy would be: 94.85 2020-02-01 20:54:03.568: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:54:03.625: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.7, 97.77, 97.84, 97.93, 98.01, 98.07, 98.14, 98.21, 98.27, 98.34, 98.41, 98.46, 98.52, 98.56, 98.61, 98.67, 98.73, 98.77, 98.81, 98.86, 98.89, 98.93, 98.96, 99.0, 99.04, 99.07, 99.11, 99.15, 99.18, 99.21, 99.22, 99.24, 99.27, 99.29, 99.32, 99.35, 99.38, 99.41, 99.45, 99.48, 99.5, 99.52, 99.55, 99.57] 2020-02-01 20:54:03.625: INFO @evaluate_confidence: Dropped ratios are: [12.02, 12.48, 13.03, 13.55, 14.02, 14.48, 14.99, 15.49, 15.98, 16.48, 16.98, 17.45, 17.89, 18.37, 18.83, 19.34, 19.8, 20.32, 20.83, 21.32, 21.84, 22.34, 22.81, 23.33, 23.85, 24.34, 24.85, 25.4, 25.96, 26.43, 26.95, 27.52, 28.11, 28.68, 29.29, 29.91, 30.54, 31.13, 31.84, 32.49, 33.19, 33.93, 34.74, 35.51] 2020-02-01 20:54:03.674: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:54:03.674: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:54:03.674: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:54:03.675: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:54:03.807: INFO @evaluate_confidence: Previous accuracy would be: 56.82 2020-02-01 20:54:03.807: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:54:03.822: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.28, 65.86, 66.57, 67.25, 67.9, 68.56, 69.45] 2020-02-01 20:54:03.822: INFO @evaluate_confidence: Dropped ratios are: [41.81, 44.78, 47.82, 50.86, 53.88, 56.71, 59.42] 2020-02-01 20:54:03.829: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:54:03.829: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.23 2020-02-01 20:54:03.829: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.22 2020-02-01 20:54:03.830: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.25 2020-02-01 20:54:03.932: INFO @evaluate_confidence: Previous accuracy would be: 91.76 2020-02-01 20:54:03.932: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:54:03.941: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.28, 96.37, 96.5, 96.6, 96.73, 96.83, 96.94, 97.01, 97.07, 97.2, 97.31, 97.43, 97.48, 97.51, 97.56, 97.65, 97.72, 97.79, 97.85, 97.89, 97.94, 98.03, 98.09, 98.15, 98.2, 98.24, 98.31, 98.36, 98.4, 98.44, 98.53, 98.59, 98.65, 98.69, 98.78, 98.85, 98.92, 98.95, 98.99, 99.02, 99.09] 2020-02-01 20:54:03.941: INFO @evaluate_confidence: Dropped ratios are: [13.67, 14.17, 14.63, 15.15, 15.65, 16.15, 16.57, 17.01, 17.48, 17.88, 18.3, 18.77, 19.15, 19.54, 19.96, 20.41, 20.91, 21.43, 21.84, 22.31, 22.9, 23.39, 23.9, 24.47, 25.02, 25.59, 26.14, 26.63, 27.2, 27.8, 28.48, 29.03, 29.67, 30.33, 30.98, 31.63, 32.31, 33.13, 34.01, 34.75, 35.61] 2020-02-01 20:54:03.948: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:54:03.948: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:54:03.948: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 20:54:03.948: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 20:54:04.072: INFO @evaluate_confidence: Previous accuracy would be: 51.90 2020-02-01 20:54:04.072: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:54:04.073: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.63, 58.06, 58.5, 58.66] 2020-02-01 20:54:04.073: INFO @evaluate_confidence: Dropped ratios are: [44.49, 48.61, 52.55, 56.47] 2020-02-01 20:54:04.124: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:54:11.102: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:54:11.184: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:54:11.642: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:54:11.887: INFO @decay_lr : LR updated to `5.99728e-05` 2020-02-01 20:54:11.888: INFO @log_profile : T train: 128.787896 2020-02-01 20:54:11.888: INFO @log_profile : T valid: 5.439590 2020-02-01 20:54:11.888: INFO @log_profile : T read data: 2.778561 2020-02-01 20:54:11.888: INFO @log_profile : T hooks: 13.175556 2020-02-01 20:54:11.888: INFO @main_loop : Epoch 102 done 2020-02-01 20:54:11.888: INFO @main_loop : Training epoch 103 2020-02-01 20:56:28.860: INFO @log_variables: train loss nanmean: 0.729809 2020-02-01 20:56:28.860: INFO @log_variables: train age_loss mean: 5.319120 2020-02-01 20:56:28.860: INFO @log_variables: train gender_loss mean: 0.134622 2020-02-01 20:56:28.860: INFO @log_variables: train age_mae mean: 5.795587 2020-02-01 20:56:28.860: INFO @log_variables: train gender_accuracy mean: 0.946307 2020-02-01 20:56:28.860: INFO @log_variables: train gender_confidence/loss nanmean: 0.054386 2020-02-01 20:56:28.860: INFO @log_variables: train gender_confidence/accuracy mean: 0.845938 2020-02-01 20:56:28.860: INFO @log_variables: train age_confidence/loss mean: 0.069484 2020-02-01 20:56:28.860: INFO @log_variables: train age_confidence/accuracy mean: 0.608254 2020-02-01 20:56:28.860: INFO @log_variables: valid loss nanmean: 0.826218 2020-02-01 20:56:28.861: INFO @log_variables: valid age_loss mean: 5.752563 2020-02-01 20:56:28.861: INFO @log_variables: valid gender_loss mean: 0.195556 2020-02-01 20:56:28.861: INFO @log_variables: valid age_mae mean: 6.232256 2020-02-01 20:56:28.861: INFO @log_variables: valid gender_accuracy mean: 0.923068 2020-02-01 20:56:28.861: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055325 2020-02-01 20:56:28.861: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873802 2020-02-01 20:56:28.861: INFO @log_variables: valid age_confidence/loss mean: 0.070155 2020-02-01 20:56:28.861: INFO @log_variables: valid age_confidence/accuracy mean: 0.557685 2020-02-01 20:56:28.861: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:56:28.868: INFO @metrics_hook: train age_mae: 5.796 +-0.033 (110592) 2020-02-01 20:56:28.874: INFO @metrics_hook: train gender_accuracy: 0.946 +-0.001 (110592) 2020-02-01 20:56:31.608: INFO @metrics_hook: valid age_mae: 6.232 +-0.088 (17639) 2020-02-01 20:56:31.610: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 20:56:33.242: INFO @decay_lr : LR updated to `5.9672937e-05` 2020-02-01 20:56:33.556: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 20:56:33.559: INFO @log_profile : T train: 128.700306 2020-02-01 20:56:33.559: INFO @log_profile : T valid: 5.716172 2020-02-01 20:56:33.559: INFO @log_profile : T read data: 1.861360 2020-02-01 20:56:33.559: INFO @log_profile : T hooks: 5.316649 2020-02-01 20:56:33.559: INFO @main_loop : Epoch 103 done 2020-02-01 20:56:33.559: INFO @main_loop : Training epoch 104 2020-02-01 20:58:45.298: INFO @log_variables: train loss nanmean: 0.725617 2020-02-01 20:58:45.298: INFO @log_variables: train age_loss mean: 5.280516 2020-02-01 20:58:45.298: INFO @log_variables: train gender_loss mean: 0.132991 2020-02-01 20:58:45.298: INFO @log_variables: train age_mae mean: 5.756648 2020-02-01 20:58:45.299: INFO @log_variables: train gender_accuracy mean: 0.947668 2020-02-01 20:58:45.299: INFO @log_variables: train gender_confidence/loss nanmean: 0.054860 2020-02-01 20:58:45.299: INFO @log_variables: train gender_confidence/accuracy mean: 0.846972 2020-02-01 20:58:45.299: INFO @log_variables: train age_confidence/loss mean: 0.069810 2020-02-01 20:58:45.299: INFO @log_variables: train age_confidence/accuracy mean: 0.606295 2020-02-01 20:58:45.299: INFO @log_variables: valid loss nanmean: 0.833205 2020-02-01 20:58:45.299: INFO @log_variables: valid age_loss mean: 5.848388 2020-02-01 20:58:45.299: INFO @log_variables: valid gender_loss mean: 0.195586 2020-02-01 20:58:45.299: INFO @log_variables: valid age_mae mean: 6.328133 2020-02-01 20:58:45.299: INFO @log_variables: valid gender_accuracy mean: 0.920971 2020-02-01 20:58:45.299: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053603 2020-02-01 20:58:45.299: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867226 2020-02-01 20:58:45.299: INFO @log_variables: valid age_confidence/loss mean: 0.070126 2020-02-01 20:58:45.299: INFO @log_variables: valid age_confidence/accuracy mean: 0.552299 2020-02-01 20:58:45.299: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 20:58:45.307: INFO @metrics_hook: train age_mae: 5.757 +-0.033 (110372) 2020-02-01 20:58:45.314: INFO @metrics_hook: train gender_accuracy: 0.948 +-0.001 (110372) 2020-02-01 20:58:48.042: INFO @metrics_hook: valid age_mae: 6.328 +-0.091 (17639) 2020-02-01 20:58:48.044: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 20:58:49.497: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:58:49.497: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 20:58:49.498: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 20:58:49.498: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 20:58:49.628: INFO @evaluate_confidence: Previous accuracy would be: 94.77 2020-02-01 20:58:49.629: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 20:58:49.688: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.75, 97.83, 97.89, 97.97, 98.04, 98.11, 98.17, 98.24, 98.3, 98.37, 98.43, 98.47, 98.52, 98.57, 98.61, 98.66, 98.71, 98.75, 98.78, 98.83, 98.88, 98.91, 98.96, 99.01, 99.06, 99.09, 99.14, 99.17, 99.2, 99.23, 99.24, 99.27, 99.31, 99.34, 99.37, 99.39, 99.42, 99.44, 99.46, 99.48, 99.5, 99.52, 99.53, 99.55] 2020-02-01 20:58:49.688: INFO @evaluate_confidence: Dropped ratios are: [12.05, 12.55, 13.06, 13.55, 14.02, 14.48, 14.95, 15.4, 15.86, 16.3, 16.79, 17.23, 17.74, 18.2, 18.68, 19.16, 19.63, 20.09, 20.55, 21.04, 21.55, 22.06, 22.58, 23.12, 23.64, 24.15, 24.67, 25.21, 25.77, 26.31, 26.85, 27.41, 27.96, 28.54, 29.1, 29.68, 30.31, 30.96, 31.68, 32.37, 33.06, 33.78, 34.54, 35.32] 2020-02-01 20:58:49.738: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:58:49.738: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.56 +- 0.14 2020-02-01 20:58:49.738: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 20:58:49.739: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 20:58:49.873: INFO @evaluate_confidence: Previous accuracy would be: 56.74 2020-02-01 20:58:49.873: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 20:58:49.888: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.12, 65.8, 66.44, 67.13, 67.96, 68.62, 69.4] 2020-02-01 20:58:49.888: INFO @evaluate_confidence: Dropped ratios are: [41.91, 45.01, 48.08, 51.09, 54.0, 56.84, 59.54] 2020-02-01 20:58:49.896: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 20:58:49.896: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.22 2020-02-01 20:58:49.896: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.22 2020-02-01 20:58:49.896: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 20:58:49.996: INFO @evaluate_confidence: Previous accuracy would be: 92.10 2020-02-01 20:58:49.997: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 20:58:50.005: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.47, 96.55, 96.63, 96.73, 96.81, 96.93, 97.0, 97.13, 97.2, 97.24, 97.36, 97.43, 97.48, 97.53, 97.64, 97.73, 97.77, 97.82, 97.9, 97.98, 98.04, 98.09, 98.13, 98.2, 98.26, 98.31, 98.4, 98.45, 98.52, 98.58, 98.63, 98.69, 98.74, 98.82, 98.85, 98.89, 98.91, 98.96, 99.01, 99.1, 99.14] 2020-02-01 20:58:50.005: INFO @evaluate_confidence: Dropped ratios are: [13.39, 13.74, 14.15, 14.58, 14.96, 15.36, 15.84, 16.37, 16.72, 17.14, 17.55, 18.02, 18.45, 18.96, 19.42, 19.92, 20.35, 20.78, 21.32, 21.74, 22.21, 22.71, 23.19, 23.73, 24.29, 24.85, 25.4, 26.07, 26.67, 27.21, 27.85, 28.39, 29.13, 29.78, 30.42, 31.12, 31.8, 32.64, 33.37, 34.21, 35.08] 2020-02-01 20:58:50.013: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 20:58:50.013: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 20:58:50.013: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 20:58:50.013: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 20:58:50.138: INFO @evaluate_confidence: Previous accuracy would be: 52.44 2020-02-01 20:58:50.138: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 20:58:50.139: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.93, 58.27, 58.68, 59.02] 2020-02-01 20:58:50.139: INFO @evaluate_confidence: Dropped ratios are: [46.25, 51.08, 55.83, 60.35] 2020-02-01 20:58:50.190: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 20:58:50.872: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:58:50.955: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 20:58:51.430: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 20:58:51.684: INFO @decay_lr : LR updated to `5.937457e-05` 2020-02-01 20:58:51.685: INFO @log_profile : T train: 122.692897 2020-02-01 20:58:51.685: INFO @log_profile : T valid: 5.513351 2020-02-01 20:58:51.685: INFO @log_profile : T read data: 2.841891 2020-02-01 20:58:51.686: INFO @log_profile : T hooks: 7.001151 2020-02-01 20:58:51.686: INFO @main_loop : Epoch 104 done 2020-02-01 20:58:51.686: INFO @main_loop : Training epoch 105 2020-02-01 21:01:10.344: INFO @log_variables: train loss nanmean: 0.722163 2020-02-01 21:01:10.344: INFO @log_variables: train age_loss mean: 5.261549 2020-02-01 21:01:10.344: INFO @log_variables: train gender_loss mean: 0.130642 2020-02-01 21:01:10.344: INFO @log_variables: train age_mae mean: 5.737714 2020-02-01 21:01:10.344: INFO @log_variables: train gender_accuracy mean: 0.948456 2020-02-01 21:01:10.344: INFO @log_variables: train gender_confidence/loss nanmean: 0.055104 2020-02-01 21:01:10.344: INFO @log_variables: train gender_confidence/accuracy mean: 0.847987 2020-02-01 21:01:10.344: INFO @log_variables: train age_confidence/loss mean: 0.069971 2020-02-01 21:01:10.344: INFO @log_variables: train age_confidence/accuracy mean: 0.604121 2020-02-01 21:01:10.344: INFO @log_variables: valid loss nanmean: 0.821707 2020-02-01 21:01:10.344: INFO @log_variables: valid age_loss mean: 5.778183 2020-02-01 21:01:10.344: INFO @log_variables: valid gender_loss mean: 0.191397 2020-02-01 21:01:10.344: INFO @log_variables: valid age_mae mean: 6.257520 2020-02-01 21:01:10.345: INFO @log_variables: valid gender_accuracy mean: 0.923068 2020-02-01 21:01:10.345: INFO @log_variables: valid gender_confidence/loss nanmean: 0.051908 2020-02-01 21:01:10.345: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868190 2020-02-01 21:01:10.345: INFO @log_variables: valid age_confidence/loss mean: 0.070512 2020-02-01 21:01:10.345: INFO @log_variables: valid age_confidence/accuracy mean: 0.554113 2020-02-01 21:01:10.345: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:01:10.352: INFO @metrics_hook: train age_mae: 5.738 +-0.033 (110372) 2020-02-01 21:01:10.359: INFO @metrics_hook: train gender_accuracy: 0.948 +-0.001 (110372) 2020-02-01 21:01:13.127: INFO @metrics_hook: valid age_mae: 6.258 +-0.090 (17639) 2020-02-01 21:01:13.129: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 21:01:14.802: INFO @decay_lr : LR updated to `5.9077698e-05` 2020-02-01 21:01:14.803: INFO @log_profile : T train: 129.628530 2020-02-01 21:01:14.803: INFO @log_profile : T valid: 5.532629 2020-02-01 21:01:14.803: INFO @log_profile : T read data: 2.807131 2020-02-01 21:01:14.804: INFO @log_profile : T hooks: 5.073263 2020-02-01 21:01:14.804: INFO @main_loop : Epoch 105 done 2020-02-01 21:01:14.804: INFO @main_loop : Training epoch 106 2020-02-01 21:03:24.784: INFO @log_variables: train loss nanmean: 0.720233 2020-02-01 21:03:24.784: INFO @log_variables: train age_loss mean: 5.239399 2020-02-01 21:03:24.784: INFO @log_variables: train gender_loss mean: 0.130746 2020-02-01 21:03:24.784: INFO @log_variables: train age_mae mean: 5.715426 2020-02-01 21:03:24.785: INFO @log_variables: train gender_accuracy mean: 0.948830 2020-02-01 21:03:24.785: INFO @log_variables: train gender_confidence/loss nanmean: 0.055146 2020-02-01 21:03:24.785: INFO @log_variables: train gender_confidence/accuracy mean: 0.846725 2020-02-01 21:03:24.785: INFO @log_variables: train age_confidence/loss mean: 0.069917 2020-02-01 21:03:24.785: INFO @log_variables: train age_confidence/accuracy mean: 0.606174 2020-02-01 21:03:24.785: INFO @log_variables: valid loss nanmean: 0.842580 2020-02-01 21:03:24.785: INFO @log_variables: valid age_loss mean: 5.813190 2020-02-01 21:03:24.785: INFO @log_variables: valid gender_loss mean: 0.207735 2020-02-01 21:03:24.785: INFO @log_variables: valid age_mae mean: 6.292787 2020-02-01 21:03:24.785: INFO @log_variables: valid gender_accuracy mean: 0.918703 2020-02-01 21:03:24.785: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055203 2020-02-01 21:03:24.785: INFO @log_variables: valid gender_confidence/accuracy mean: 0.881286 2020-02-01 21:03:24.785: INFO @log_variables: valid age_confidence/loss mean: 0.070055 2020-02-01 21:03:24.785: INFO @log_variables: valid age_confidence/accuracy mean: 0.566415 2020-02-01 21:03:24.785: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:03:24.793: INFO @metrics_hook: train age_mae: 5.715 +-0.033 (110592) 2020-02-01 21:03:24.800: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110592) 2020-02-01 21:03:27.573: INFO @metrics_hook: valid age_mae: 6.293 +-0.090 (17639) 2020-02-01 21:03:27.574: INFO @metrics_hook: valid gender_accuracy: 0.919 +-0.004 (17639) 2020-02-01 21:03:29.047: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:03:29.047: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 21:03:29.048: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.37 +- 0.21 2020-02-01 21:03:29.048: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:03:29.177: INFO @evaluate_confidence: Previous accuracy would be: 94.88 2020-02-01 21:03:29.178: INFO @evaluate_confidence: Possible optimal thresholds are: [0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 21:03:29.236: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.76, 97.84, 97.92, 98.0, 98.06, 98.11, 98.17, 98.22, 98.3, 98.36, 98.4, 98.45, 98.51, 98.56, 98.62, 98.66, 98.69, 98.74, 98.79, 98.82, 98.87, 98.91, 98.94, 98.97, 99.01, 99.04, 99.08, 99.12, 99.14, 99.17, 99.21, 99.25, 99.28, 99.31, 99.34, 99.36, 99.37, 99.4, 99.42, 99.45, 99.48, 99.5, 99.53, 99.54] 2020-02-01 21:03:29.236: INFO @evaluate_confidence: Dropped ratios are: [11.84, 12.31, 12.79, 13.34, 13.84, 14.31, 14.75, 15.25, 15.73, 16.19, 16.64, 17.12, 17.58, 18.08, 18.57, 19.03, 19.51, 20.03, 20.54, 21.0, 21.49, 22.0, 22.49, 22.94, 23.46, 23.97, 24.48, 25.03, 25.58, 26.12, 26.69, 27.27, 27.87, 28.48, 29.12, 29.71, 30.31, 30.97, 31.67, 32.41, 33.14, 33.89, 34.66, 35.47] 2020-02-01 21:03:29.286: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:03:29.286: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:03:29.287: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:03:29.287: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 21:03:29.423: INFO @evaluate_confidence: Previous accuracy would be: 57.10 2020-02-01 21:03:29.423: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:03:29.439: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.39, 66.03, 66.68, 67.39, 68.18, 69.01, 69.8, 70.6] 2020-02-01 21:03:29.440: INFO @evaluate_confidence: Dropped ratios are: [41.57, 44.74, 47.83, 50.8, 53.83, 56.71, 59.45, 62.16] 2020-02-01 21:03:29.447: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:03:29.447: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.21 2020-02-01 21:03:29.447: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.23 2020-02-01 21:03:29.448: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.23 2020-02-01 21:03:29.550: INFO @evaluate_confidence: Previous accuracy would be: 91.87 2020-02-01 21:03:29.550: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 21:03:29.559: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.09, 96.23, 96.31, 96.37, 96.46, 96.56, 96.66, 96.73, 96.85, 96.94, 97.03, 97.13, 97.19, 97.23, 97.33, 97.39, 97.5, 97.57, 97.66, 97.73, 97.81, 97.83, 97.93, 97.98, 98.04, 98.1, 98.17, 98.21, 98.26, 98.3, 98.34, 98.41, 98.49, 98.53, 98.56, 98.62, 98.67, 98.77, 98.85, 98.89] 2020-02-01 21:03:29.559: INFO @evaluate_confidence: Dropped ratios are: [12.5, 13.02, 13.36, 13.73, 14.12, 14.56, 15.01, 15.44, 15.87, 16.3, 16.82, 17.21, 17.6, 17.99, 18.51, 18.93, 19.28, 19.76, 20.24, 20.69, 21.14, 21.66, 22.12, 22.67, 23.23, 23.74, 24.3, 24.91, 25.46, 26.04, 26.63, 27.41, 28.09, 28.71, 29.47, 30.08, 30.91, 31.69, 32.62, 33.55] 2020-02-01 21:03:29.566: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:03:29.567: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:03:29.567: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 21:03:29.567: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.11 2020-02-01 21:03:29.693: INFO @evaluate_confidence: Previous accuracy would be: 52.92 2020-02-01 21:03:29.693: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 21:03:29.695: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.44, 59.84, 60.21, 60.31] 2020-02-01 21:03:29.695: INFO @evaluate_confidence: Dropped ratios are: [43.66, 48.43, 53.21, 58.05] 2020-02-01 21:03:29.748: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:03:30.443: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:03:30.528: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:03:30.977: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:03:31.225: INFO @decay_lr : LR updated to `5.878231e-05` 2020-02-01 21:03:31.226: INFO @log_profile : T train: 121.907188 2020-02-01 21:03:31.226: INFO @log_profile : T valid: 5.518829 2020-02-01 21:03:31.226: INFO @log_profile : T read data: 1.884910 2020-02-01 21:03:31.226: INFO @log_profile : T hooks: 7.032689 2020-02-01 21:03:31.226: INFO @main_loop : Epoch 106 done 2020-02-01 21:03:31.226: INFO @main_loop : Training epoch 107 2020-02-01 21:05:42.047: INFO @log_variables: train loss nanmean: 0.720451 2020-02-01 21:05:42.047: INFO @log_variables: train age_loss mean: 5.253423 2020-02-01 21:05:42.047: INFO @log_variables: train gender_loss mean: 0.131039 2020-02-01 21:05:42.047: INFO @log_variables: train age_mae mean: 5.729626 2020-02-01 21:05:42.047: INFO @log_variables: train gender_accuracy mean: 0.947595 2020-02-01 21:05:42.047: INFO @log_variables: train gender_confidence/loss nanmean: 0.053907 2020-02-01 21:05:42.047: INFO @log_variables: train gender_confidence/accuracy mean: 0.846881 2020-02-01 21:05:42.047: INFO @log_variables: train age_confidence/loss mean: 0.069834 2020-02-01 21:05:42.047: INFO @log_variables: train age_confidence/accuracy mean: 0.611523 2020-02-01 21:05:42.047: INFO @log_variables: valid loss nanmean: 0.825839 2020-02-01 21:05:42.047: INFO @log_variables: valid age_loss mean: 5.781447 2020-02-01 21:05:42.047: INFO @log_variables: valid gender_loss mean: 0.193572 2020-02-01 21:05:42.047: INFO @log_variables: valid age_mae mean: 6.261646 2020-02-01 21:05:42.047: INFO @log_variables: valid gender_accuracy mean: 0.924486 2020-02-01 21:05:42.047: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053637 2020-02-01 21:05:42.047: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872612 2020-02-01 21:05:42.048: INFO @log_variables: valid age_confidence/loss mean: 0.070641 2020-02-01 21:05:42.048: INFO @log_variables: valid age_confidence/accuracy mean: 0.550882 2020-02-01 21:05:42.048: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:05:42.055: INFO @metrics_hook: train age_mae: 5.730 +-0.033 (110372) 2020-02-01 21:05:42.062: INFO @metrics_hook: train gender_accuracy: 0.948 +-0.001 (110372) 2020-02-01 21:05:44.836: INFO @metrics_hook: valid age_mae: 6.262 +-0.090 (17639) 2020-02-01 21:05:44.838: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 21:05:46.511: INFO @decay_lr : LR updated to `5.8488396e-05` 2020-02-01 21:05:46.832: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 21:05:46.836: INFO @log_profile : T train: 121.781063 2020-02-01 21:05:46.836: INFO @log_profile : T valid: 5.461757 2020-02-01 21:05:46.836: INFO @log_profile : T read data: 2.883485 2020-02-01 21:05:46.836: INFO @log_profile : T hooks: 5.406276 2020-02-01 21:05:46.836: INFO @main_loop : Epoch 107 done 2020-02-01 21:05:46.836: INFO @main_loop : Training epoch 108 2020-02-01 21:07:57.369: INFO @log_variables: train loss nanmean: 0.720110 2020-02-01 21:07:57.369: INFO @log_variables: train age_loss mean: 5.235104 2020-02-01 21:07:57.369: INFO @log_variables: train gender_loss mean: 0.131515 2020-02-01 21:07:57.369: INFO @log_variables: train age_mae mean: 5.711148 2020-02-01 21:07:57.369: INFO @log_variables: train gender_accuracy mean: 0.948121 2020-02-01 21:07:57.369: INFO @log_variables: train gender_confidence/loss nanmean: 0.054763 2020-02-01 21:07:57.369: INFO @log_variables: train gender_confidence/accuracy mean: 0.846909 2020-02-01 21:07:57.369: INFO @log_variables: train age_confidence/loss mean: 0.069869 2020-02-01 21:07:57.369: INFO @log_variables: train age_confidence/accuracy mean: 0.609484 2020-02-01 21:07:57.369: INFO @log_variables: valid loss nanmean: 0.844058 2020-02-01 21:07:57.370: INFO @log_variables: valid age_loss mean: 5.847793 2020-02-01 21:07:57.370: INFO @log_variables: valid gender_loss mean: 0.206628 2020-02-01 21:07:57.370: INFO @log_variables: valid age_mae mean: 6.327667 2020-02-01 21:07:57.370: INFO @log_variables: valid gender_accuracy mean: 0.921538 2020-02-01 21:07:57.370: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054882 2020-02-01 21:07:57.370: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873292 2020-02-01 21:07:57.370: INFO @log_variables: valid age_confidence/loss mean: 0.069715 2020-02-01 21:07:57.370: INFO @log_variables: valid age_confidence/accuracy mean: 0.566699 2020-02-01 21:07:57.370: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:07:57.377: INFO @metrics_hook: train age_mae: 5.711 +-0.033 (110372) 2020-02-01 21:07:57.384: INFO @metrics_hook: train gender_accuracy: 0.948 +-0.001 (110372) 2020-02-01 21:08:00.185: INFO @metrics_hook: valid age_mae: 6.328 +-0.090 (17639) 2020-02-01 21:08:00.186: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 21:08:01.673: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:08:01.673: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 21:08:01.673: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.21 2020-02-01 21:08:01.673: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:08:01.809: INFO @evaluate_confidence: Previous accuracy would be: 94.81 2020-02-01 21:08:01.809: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 21:08:01.871: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.69, 97.8, 97.88, 97.97, 98.04, 98.11, 98.16, 98.22, 98.29, 98.36, 98.41, 98.46, 98.5, 98.55, 98.6, 98.66, 98.71, 98.76, 98.8, 98.84, 98.87, 98.92, 98.95, 98.99, 99.02, 99.05, 99.09, 99.12, 99.16, 99.19, 99.21, 99.23, 99.26, 99.28, 99.31, 99.34, 99.37, 99.39, 99.42, 99.44, 99.47, 99.49, 99.52, 99.53, 99.54] 2020-02-01 21:08:01.871: INFO @evaluate_confidence: Dropped ratios are: [11.68, 12.15, 12.64, 13.08, 13.51, 13.96, 14.39, 14.87, 15.35, 15.84, 16.31, 16.81, 17.27, 17.75, 18.21, 18.68, 19.15, 19.61, 20.06, 20.53, 21.03, 21.48, 21.91, 22.38, 22.87, 23.35, 23.84, 24.32, 24.84, 25.4, 25.94, 26.48, 27.02, 27.55, 28.11, 28.71, 29.27, 29.87, 30.52, 31.18, 31.86, 32.57, 33.28, 34.03, 34.84] 2020-02-01 21:08:01.921: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:08:01.921: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:08:01.921: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:08:01.922: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.14 2020-02-01 21:08:02.061: INFO @evaluate_confidence: Previous accuracy would be: 57.07 2020-02-01 21:08:02.061: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:08:02.078: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.49, 66.15, 66.89, 67.63, 68.34, 68.97, 69.81, 70.68] 2020-02-01 21:08:02.078: INFO @evaluate_confidence: Dropped ratios are: [41.06, 44.22, 47.38, 50.56, 53.61, 56.53, 59.29, 61.94] 2020-02-01 21:08:02.086: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:08:02.086: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:08:02.086: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.23 2020-02-01 21:08:02.086: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 21:08:02.191: INFO @evaluate_confidence: Previous accuracy would be: 92.15 2020-02-01 21:08:02.192: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:08:02.200: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.22, 96.3, 96.37, 96.48, 96.56, 96.68, 96.79, 96.87, 96.96, 97.01, 97.13, 97.21, 97.29, 97.39, 97.47, 97.56, 97.65, 97.72, 97.79, 97.83, 97.88, 97.91, 97.95, 98.01, 98.06, 98.11, 98.15, 98.19, 98.27, 98.3, 98.37, 98.44, 98.48, 98.57, 98.61, 98.66, 98.72, 98.79, 98.85, 98.9, 98.94] 2020-02-01 21:08:02.200: INFO @evaluate_confidence: Dropped ratios are: [12.86, 13.23, 13.59, 13.98, 14.41, 14.87, 15.32, 15.78, 16.21, 16.58, 16.97, 17.41, 17.79, 18.21, 18.65, 19.11, 19.55, 20.02, 20.57, 21.08, 21.54, 22.01, 22.5, 22.95, 23.56, 24.08, 24.58, 25.04, 25.52, 26.11, 26.74, 27.33, 27.9, 28.62, 29.25, 29.98, 30.65, 31.46, 32.27, 33.15, 33.95] 2020-02-01 21:08:02.208: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:08:02.208: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 21:08:02.208: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 21:08:02.209: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.11 2020-02-01 21:08:02.337: INFO @evaluate_confidence: Previous accuracy would be: 52.25 2020-02-01 21:08:02.337: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 21:08:02.338: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.66, 58.97, 59.25, 59.37] 2020-02-01 21:08:02.338: INFO @evaluate_confidence: Dropped ratios are: [45.65, 50.33, 54.78, 59.27] 2020-02-01 21:08:02.390: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:08:03.089: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:08:03.175: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:08:03.643: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:08:03.907: INFO @decay_lr : LR updated to `5.8195954e-05` 2020-02-01 21:08:03.908: INFO @log_profile : T train: 121.478967 2020-02-01 21:08:03.908: INFO @log_profile : T valid: 5.474249 2020-02-01 21:08:03.908: INFO @log_profile : T read data: 2.914908 2020-02-01 21:08:03.909: INFO @log_profile : T hooks: 7.126342 2020-02-01 21:08:03.909: INFO @main_loop : Epoch 108 done 2020-02-01 21:08:03.909: INFO @main_loop : Training epoch 109 2020-02-01 21:10:13.926: INFO @log_variables: train loss nanmean: 0.717160 2020-02-01 21:10:13.926: INFO @log_variables: train age_loss mean: 5.241360 2020-02-01 21:10:13.926: INFO @log_variables: train gender_loss mean: 0.129060 2020-02-01 21:10:13.926: INFO @log_variables: train age_mae mean: 5.718088 2020-02-01 21:10:13.927: INFO @log_variables: train gender_accuracy mean: 0.948559 2020-02-01 21:10:13.927: INFO @log_variables: train gender_confidence/loss nanmean: 0.053485 2020-02-01 21:10:13.927: INFO @log_variables: train gender_confidence/accuracy mean: 0.847385 2020-02-01 21:10:13.927: INFO @log_variables: train age_confidence/loss mean: 0.069860 2020-02-01 21:10:13.927: INFO @log_variables: train age_confidence/accuracy mean: 0.604429 2020-02-01 21:10:13.927: INFO @log_variables: valid loss nanmean: 0.850171 2020-02-01 21:10:13.927: INFO @log_variables: valid age_loss mean: 5.886659 2020-02-01 21:10:13.927: INFO @log_variables: valid gender_loss mean: 0.209599 2020-02-01 21:10:13.927: INFO @log_variables: valid age_mae mean: 6.368063 2020-02-01 21:10:13.927: INFO @log_variables: valid gender_accuracy mean: 0.917286 2020-02-01 21:10:13.927: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055157 2020-02-01 21:10:13.927: INFO @log_variables: valid gender_confidence/accuracy mean: 0.863768 2020-02-01 21:10:13.927: INFO @log_variables: valid age_confidence/loss mean: 0.069318 2020-02-01 21:10:13.927: INFO @log_variables: valid age_confidence/accuracy mean: 0.562900 2020-02-01 21:10:13.927: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:10:13.935: INFO @metrics_hook: train age_mae: 5.718 +-0.033 (110592) 2020-02-01 21:10:13.941: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110592) 2020-02-01 21:10:16.750: INFO @metrics_hook: valid age_mae: 6.368 +-0.090 (17639) 2020-02-01 21:10:16.751: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 21:10:18.411: INFO @decay_lr : LR updated to `5.7904974e-05` 2020-02-01 21:10:18.412: INFO @log_profile : T train: 121.942386 2020-02-01 21:10:18.412: INFO @log_profile : T valid: 5.467525 2020-02-01 21:10:18.412: INFO @log_profile : T read data: 1.920688 2020-02-01 21:10:18.413: INFO @log_profile : T hooks: 5.097310 2020-02-01 21:10:18.413: INFO @main_loop : Epoch 109 done 2020-02-01 21:10:18.413: INFO @main_loop : Training epoch 110 2020-02-01 21:12:30.789: INFO @log_variables: train loss nanmean: 0.718160 2020-02-01 21:12:30.789: INFO @log_variables: train age_loss mean: 5.244355 2020-02-01 21:12:30.789: INFO @log_variables: train gender_loss mean: 0.129227 2020-02-01 21:12:30.789: INFO @log_variables: train age_mae mean: 5.720185 2020-02-01 21:12:30.789: INFO @log_variables: train gender_accuracy mean: 0.948755 2020-02-01 21:12:30.789: INFO @log_variables: train gender_confidence/loss nanmean: 0.053963 2020-02-01 21:12:30.789: INFO @log_variables: train gender_confidence/accuracy mean: 0.846818 2020-02-01 21:12:30.789: INFO @log_variables: train age_confidence/loss mean: 0.069958 2020-02-01 21:12:30.789: INFO @log_variables: train age_confidence/accuracy mean: 0.607917 2020-02-01 21:12:30.789: INFO @log_variables: valid loss nanmean: 0.828333 2020-02-01 21:12:30.789: INFO @log_variables: valid age_loss mean: 5.780851 2020-02-01 21:12:30.789: INFO @log_variables: valid gender_loss mean: 0.196967 2020-02-01 21:12:30.789: INFO @log_variables: valid age_mae mean: 6.260375 2020-02-01 21:12:30.789: INFO @log_variables: valid gender_accuracy mean: 0.923748 2020-02-01 21:12:30.790: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053567 2020-02-01 21:12:30.790: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871875 2020-02-01 21:12:30.790: INFO @log_variables: valid age_confidence/loss mean: 0.070172 2020-02-01 21:12:30.790: INFO @log_variables: valid age_confidence/accuracy mean: 0.561540 2020-02-01 21:12:30.790: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:12:30.797: INFO @metrics_hook: train age_mae: 5.720 +-0.033 (110372) 2020-02-01 21:12:30.804: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110372) 2020-02-01 21:12:33.516: INFO @metrics_hook: valid age_mae: 6.260 +-0.089 (17639) 2020-02-01 21:12:33.517: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 21:12:34.986: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:12:34.986: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:12:34.986: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.20 2020-02-01 21:12:34.986: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:12:35.118: INFO @evaluate_confidence: Previous accuracy would be: 94.88 2020-02-01 21:12:35.118: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:12:35.179: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.7, 97.77, 97.83, 97.91, 97.99, 98.07, 98.14, 98.2, 98.26, 98.33, 98.39, 98.45, 98.5, 98.55, 98.6, 98.64, 98.69, 98.74, 98.8, 98.85, 98.9, 98.95, 99.0, 99.04, 99.06, 99.1, 99.13, 99.16, 99.18, 99.2, 99.23, 99.27, 99.3, 99.34, 99.36, 99.39, 99.42, 99.44, 99.46, 99.48, 99.51, 99.56, 99.58, 99.6, 99.62, 99.64] 2020-02-01 21:12:35.179: INFO @evaluate_confidence: Dropped ratios are: [11.75, 12.21, 12.65, 13.11, 13.57, 14.04, 14.5, 14.94, 15.38, 15.85, 16.33, 16.79, 17.26, 17.67, 18.14, 18.61, 19.08, 19.53, 20.02, 20.49, 20.94, 21.43, 21.91, 22.34, 22.82, 23.3, 23.77, 24.28, 24.75, 25.27, 25.77, 26.31, 26.84, 27.39, 27.99, 28.57, 29.13, 29.74, 30.41, 31.04, 31.73, 32.42, 33.1, 33.82, 34.55, 35.38] 2020-02-01 21:12:35.231: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:12:35.231: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:12:35.231: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:12:35.232: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.13 2020-02-01 21:12:35.370: INFO @evaluate_confidence: Previous accuracy would be: 57.16 2020-02-01 21:12:35.370: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:12:35.387: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.45, 66.08, 66.75, 67.43, 68.21, 69.0, 69.89, 70.83] 2020-02-01 21:12:35.387: INFO @evaluate_confidence: Dropped ratios are: [41.03, 44.21, 47.47, 50.59, 53.71, 56.64, 59.44, 62.13] 2020-02-01 21:12:35.394: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:12:35.394: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:12:35.394: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.23 2020-02-01 21:12:35.395: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 21:12:35.500: INFO @evaluate_confidence: Previous accuracy would be: 92.37 2020-02-01 21:12:35.500: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:12:35.509: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.34, 96.44, 96.56, 96.67, 96.81, 96.87, 96.96, 97.02, 97.11, 97.22, 97.29, 97.38, 97.49, 97.56, 97.62, 97.67, 97.75, 97.81, 97.85, 97.95, 98.03, 98.06, 98.09, 98.13, 98.18, 98.26, 98.3, 98.35, 98.38, 98.44, 98.51, 98.59, 98.63, 98.65, 98.71, 98.74, 98.79, 98.82, 98.89, 98.98, 99.02] 2020-02-01 21:12:35.509: INFO @evaluate_confidence: Dropped ratios are: [13.0, 13.44, 13.91, 14.35, 14.78, 15.13, 15.59, 16.03, 16.5, 16.9, 17.28, 17.78, 18.16, 18.66, 19.14, 19.6, 19.99, 20.38, 20.77, 21.32, 21.74, 22.14, 22.65, 23.1, 23.64, 24.14, 24.8, 25.31, 25.82, 26.45, 27.05, 27.69, 28.3, 28.83, 29.46, 30.19, 30.93, 31.73, 32.48, 33.53, 34.49] 2020-02-01 21:12:35.517: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:12:35.517: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:12:35.517: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 21:12:35.517: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 21:12:35.649: INFO @evaluate_confidence: Previous accuracy would be: 53.00 2020-02-01 21:12:35.650: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 21:12:35.651: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.21, 59.52, 59.69, 59.88] 2020-02-01 21:12:35.651: INFO @evaluate_confidence: Dropped ratios are: [44.31, 48.54, 52.71, 57.15] 2020-02-01 21:12:35.704: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:12:36.392: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:12:36.478: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:12:36.934: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:12:37.008: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:12:37.685: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:12:37.766: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:12:37.768: INFO @evaluate_gender-age_model: groups 0 3.626771 1 4.229584 2 5.458326 3 5.658065 4 6.361914 5 6.306004 6 6.374142 7 7.161878 Name: errors, dtype: float64 2020-02-01 21:12:37.769: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:12:38.217: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:12:38.275: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:12:38.276: INFO @evaluate_gender-age_model: groups 0 6.505264 1 5.536145 2 5.511030 3 5.382254 4 7.193863 5 5.681698 6 7.724266 7 12.124997 Name: errors, dtype: float64 2020-02-01 21:12:38.456: INFO @decay_lr : LR updated to `5.761545e-05` 2020-02-01 21:12:38.457: INFO @log_profile : T train: 121.570652 2020-02-01 21:12:38.457: INFO @log_profile : T valid: 5.402451 2020-02-01 21:12:38.457: INFO @log_profile : T read data: 2.871109 2020-02-01 21:12:38.458: INFO @log_profile : T hooks: 10.122818 2020-02-01 21:12:38.458: INFO @main_loop : Epoch 110 done 2020-02-01 21:12:38.458: INFO @main_loop : Training epoch 111 2020-02-01 21:14:48.477: INFO @log_variables: train loss nanmean: 0.714809 2020-02-01 21:14:48.477: INFO @log_variables: train age_loss mean: 5.206955 2020-02-01 21:14:48.477: INFO @log_variables: train gender_loss mean: 0.128406 2020-02-01 21:14:48.477: INFO @log_variables: train age_mae mean: 5.683255 2020-02-01 21:14:48.477: INFO @log_variables: train gender_accuracy mean: 0.949562 2020-02-01 21:14:48.478: INFO @log_variables: train gender_confidence/loss nanmean: 0.054612 2020-02-01 21:14:48.478: INFO @log_variables: train gender_confidence/accuracy mean: 0.846282 2020-02-01 21:14:48.478: INFO @log_variables: train age_confidence/loss mean: 0.070105 2020-02-01 21:14:48.478: INFO @log_variables: train age_confidence/accuracy mean: 0.606906 2020-02-01 21:14:48.478: INFO @log_variables: valid loss nanmean: 0.834293 2020-02-01 21:14:48.478: INFO @log_variables: valid age_loss mean: 5.812351 2020-02-01 21:14:48.478: INFO @log_variables: valid gender_loss mean: 0.199597 2020-02-01 21:14:48.478: INFO @log_variables: valid age_mae mean: 6.292713 2020-02-01 21:14:48.478: INFO @log_variables: valid gender_accuracy mean: 0.922218 2020-02-01 21:14:48.478: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054198 2020-02-01 21:14:48.478: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871478 2020-02-01 21:14:48.478: INFO @log_variables: valid age_confidence/loss mean: 0.070248 2020-02-01 21:14:48.478: INFO @log_variables: valid age_confidence/accuracy mean: 0.566132 2020-02-01 21:14:48.478: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:14:48.486: INFO @metrics_hook: train age_mae: 5.683 +-0.033 (110592) 2020-02-01 21:14:48.493: INFO @metrics_hook: train gender_accuracy: 0.950 +-0.001 (110592) 2020-02-01 21:14:51.286: INFO @metrics_hook: valid age_mae: 6.293 +-0.090 (17639) 2020-02-01 21:14:51.287: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 21:14:52.926: INFO @decay_lr : LR updated to `5.7327372e-05` 2020-02-01 21:14:52.927: INFO @log_profile : T train: 121.990311 2020-02-01 21:14:52.927: INFO @log_profile : T valid: 5.437301 2020-02-01 21:14:52.927: INFO @log_profile : T read data: 1.871650 2020-02-01 21:14:52.927: INFO @log_profile : T hooks: 5.092480 2020-02-01 21:14:52.928: INFO @main_loop : Epoch 111 done 2020-02-01 21:14:52.928: INFO @main_loop : Training epoch 112 2020-02-01 21:17:03.388: INFO @log_variables: train loss nanmean: 0.715294 2020-02-01 21:17:03.388: INFO @log_variables: train age_loss mean: 5.204088 2020-02-01 21:17:03.389: INFO @log_variables: train gender_loss mean: 0.129337 2020-02-01 21:17:03.389: INFO @log_variables: train age_mae mean: 5.679802 2020-02-01 21:17:03.389: INFO @log_variables: train gender_accuracy mean: 0.948882 2020-02-01 21:17:03.389: INFO @log_variables: train gender_confidence/loss nanmean: 0.054377 2020-02-01 21:17:03.389: INFO @log_variables: train gender_confidence/accuracy mean: 0.846836 2020-02-01 21:17:03.389: INFO @log_variables: train age_confidence/loss mean: 0.070239 2020-02-01 21:17:03.389: INFO @log_variables: train age_confidence/accuracy mean: 0.607935 2020-02-01 21:17:03.389: INFO @log_variables: valid loss nanmean: 0.833606 2020-02-01 21:17:03.389: INFO @log_variables: valid age_loss mean: 5.815196 2020-02-01 21:17:03.389: INFO @log_variables: valid gender_loss mean: 0.200282 2020-02-01 21:17:03.389: INFO @log_variables: valid age_mae mean: 6.296081 2020-02-01 21:17:03.389: INFO @log_variables: valid gender_accuracy mean: 0.921141 2020-02-01 21:17:03.389: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052792 2020-02-01 21:17:03.389: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870628 2020-02-01 21:17:03.389: INFO @log_variables: valid age_confidence/loss mean: 0.070085 2020-02-01 21:17:03.389: INFO @log_variables: valid age_confidence/accuracy mean: 0.554907 2020-02-01 21:17:03.389: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:17:03.397: INFO @metrics_hook: train age_mae: 5.680 +-0.033 (110372) 2020-02-01 21:17:03.404: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110372) 2020-02-01 21:17:06.132: INFO @metrics_hook: valid age_mae: 6.296 +-0.090 (17639) 2020-02-01 21:17:06.133: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 21:17:07.608: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:17:07.608: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.81 +- 0.24 2020-02-01 21:17:07.608: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.21 2020-02-01 21:17:07.609: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:17:07.740: INFO @evaluate_confidence: Previous accuracy would be: 94.89 2020-02-01 21:17:07.740: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8] 2020-02-01 21:17:07.804: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.68, 97.77, 97.85, 97.93, 98.03, 98.07, 98.14, 98.2, 98.26, 98.32, 98.38, 98.45, 98.49, 98.54, 98.6, 98.64, 98.69, 98.74, 98.8, 98.85, 98.88, 98.93, 98.96, 99.0, 99.02, 99.07, 99.11, 99.15, 99.18, 99.21, 99.25, 99.28, 99.31, 99.34, 99.36, 99.38, 99.41, 99.43, 99.45, 99.48, 99.5, 99.53, 99.56, 99.58, 99.6] 2020-02-01 21:17:07.804: INFO @evaluate_confidence: Dropped ratios are: [11.63, 12.1, 12.59, 13.08, 13.56, 14.04, 14.48, 14.92, 15.35, 15.82, 16.27, 16.76, 17.19, 17.67, 18.13, 18.6, 19.06, 19.56, 20.07, 20.57, 21.02, 21.5, 21.97, 22.44, 22.9, 23.38, 23.9, 24.45, 24.97, 25.47, 26.03, 26.59, 27.15, 27.7, 28.28, 28.85, 29.43, 30.04, 30.68, 31.31, 31.98, 32.67, 33.41, 34.14, 34.9] 2020-02-01 21:17:07.854: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:17:07.855: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:17:07.855: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:17:07.855: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 21:17:07.994: INFO @evaluate_confidence: Previous accuracy would be: 57.63 2020-02-01 21:17:07.994: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:17:08.012: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.64, 66.32, 67.05, 67.73, 68.42, 69.2, 69.92, 70.79] 2020-02-01 21:17:08.012: INFO @evaluate_confidence: Dropped ratios are: [40.37, 43.57, 46.6, 49.7, 52.67, 55.58, 58.4, 61.14] 2020-02-01 21:17:08.020: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:17:08.020: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:17:08.020: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.22 2020-02-01 21:17:08.020: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 21:17:08.129: INFO @evaluate_confidence: Previous accuracy would be: 92.11 2020-02-01 21:17:08.129: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:17:08.138: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.29, 96.4, 96.49, 96.58, 96.66, 96.78, 96.86, 96.95, 97.05, 97.14, 97.28, 97.37, 97.47, 97.54, 97.61, 97.7, 97.75, 97.79, 97.9, 97.97, 98.04, 98.08, 98.18, 98.26, 98.3, 98.32, 98.4, 98.47, 98.51, 98.53, 98.61, 98.64, 98.71, 98.76, 98.85, 98.89, 98.92, 98.95, 98.98, 99.02, 99.05] 2020-02-01 21:17:08.138: INFO @evaluate_confidence: Dropped ratios are: [13.43, 13.79, 14.14, 14.52, 14.92, 15.38, 15.77, 16.1, 16.54, 17.03, 17.49, 18.02, 18.54, 19.03, 19.51, 19.97, 20.38, 20.86, 21.41, 21.86, 22.39, 22.87, 23.41, 23.94, 24.44, 24.88, 25.45, 25.9, 26.38, 26.97, 27.62, 28.24, 28.84, 29.37, 30.02, 30.82, 31.49, 32.27, 33.02, 33.85, 34.75] 2020-02-01 21:17:08.145: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:17:08.145: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:17:08.145: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 21:17:08.145: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 21:17:08.270: INFO @evaluate_confidence: Previous accuracy would be: 52.73 2020-02-01 21:17:08.270: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 21:17:08.271: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.42, 58.81, 59.09, 59.35] 2020-02-01 21:17:08.271: INFO @evaluate_confidence: Dropped ratios are: [46.02, 50.41, 54.79, 59.44] 2020-02-01 21:17:08.325: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:17:09.021: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:17:09.103: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:17:09.557: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:17:09.806: INFO @decay_lr : LR updated to `5.7040736e-05` 2020-02-01 21:17:09.807: INFO @log_profile : T train: 121.532501 2020-02-01 21:17:09.807: INFO @log_profile : T valid: 5.406604 2020-02-01 21:17:09.807: INFO @log_profile : T read data: 2.828655 2020-02-01 21:17:09.807: INFO @log_profile : T hooks: 7.034874 2020-02-01 21:17:09.807: INFO @main_loop : Epoch 112 done 2020-02-01 21:17:09.807: INFO @main_loop : Training epoch 113 2020-02-01 21:19:29.140: INFO @log_variables: train loss nanmean: 0.712416 2020-02-01 21:19:29.140: INFO @log_variables: train age_loss mean: 5.189148 2020-02-01 21:19:29.140: INFO @log_variables: train gender_loss mean: 0.128393 2020-02-01 21:19:29.141: INFO @log_variables: train age_mae mean: 5.664850 2020-02-01 21:19:29.141: INFO @log_variables: train gender_accuracy mean: 0.949054 2020-02-01 21:19:29.141: INFO @log_variables: train gender_confidence/loss nanmean: 0.053791 2020-02-01 21:19:29.141: INFO @log_variables: train gender_confidence/accuracy mean: 0.847570 2020-02-01 21:19:29.141: INFO @log_variables: train age_confidence/loss mean: 0.070164 2020-02-01 21:19:29.141: INFO @log_variables: train age_confidence/accuracy mean: 0.607029 2020-02-01 21:19:29.141: INFO @log_variables: valid loss nanmean: 0.825208 2020-02-01 21:19:29.141: INFO @log_variables: valid age_loss mean: 5.763188 2020-02-01 21:19:29.141: INFO @log_variables: valid gender_loss mean: 0.195614 2020-02-01 21:19:29.141: INFO @log_variables: valid age_mae mean: 6.243295 2020-02-01 21:19:29.141: INFO @log_variables: valid gender_accuracy mean: 0.924939 2020-02-01 21:19:29.141: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053447 2020-02-01 21:19:29.141: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875050 2020-02-01 21:19:29.141: INFO @log_variables: valid age_confidence/loss mean: 0.070004 2020-02-01 21:19:29.141: INFO @log_variables: valid age_confidence/accuracy mean: 0.577527 2020-02-01 21:19:29.141: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:19:29.149: INFO @metrics_hook: train age_mae: 5.665 +-0.033 (110372) 2020-02-01 21:19:29.156: INFO @metrics_hook: train gender_accuracy: 0.949 +-0.001 (110372) 2020-02-01 21:19:31.947: INFO @metrics_hook: valid age_mae: 6.243 +-0.089 (17639) 2020-02-01 21:19:31.948: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 21:19:33.595: INFO @decay_lr : LR updated to `5.6755533e-05` 2020-02-01 21:19:33.902: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 21:19:33.906: INFO @log_profile : T train: 130.173260 2020-02-01 21:19:33.906: INFO @log_profile : T valid: 5.670856 2020-02-01 21:19:33.906: INFO @log_profile : T read data: 2.797741 2020-02-01 21:19:33.906: INFO @log_profile : T hooks: 5.380306 2020-02-01 21:19:33.906: INFO @main_loop : Epoch 113 done 2020-02-01 21:19:33.906: INFO @main_loop : Training epoch 114 2020-02-01 21:21:43.875: INFO @log_variables: train loss nanmean: 0.707491 2020-02-01 21:21:43.876: INFO @log_variables: train age_loss mean: 5.173041 2020-02-01 21:21:43.876: INFO @log_variables: train gender_loss mean: 0.125368 2020-02-01 21:21:43.876: INFO @log_variables: train age_mae mean: 5.648499 2020-02-01 21:21:43.876: INFO @log_variables: train gender_accuracy mean: 0.950458 2020-02-01 21:21:43.876: INFO @log_variables: train gender_confidence/loss nanmean: 0.052941 2020-02-01 21:21:43.876: INFO @log_variables: train gender_confidence/accuracy mean: 0.851318 2020-02-01 21:21:43.876: INFO @log_variables: train age_confidence/loss mean: 0.070303 2020-02-01 21:21:43.876: INFO @log_variables: train age_confidence/accuracy mean: 0.606635 2020-02-01 21:21:43.876: INFO @log_variables: valid loss nanmean: 0.832867 2020-02-01 21:21:43.876: INFO @log_variables: valid age_loss mean: 5.825095 2020-02-01 21:21:43.876: INFO @log_variables: valid gender_loss mean: 0.197189 2020-02-01 21:21:43.876: INFO @log_variables: valid age_mae mean: 6.304503 2020-02-01 21:21:43.876: INFO @log_variables: valid gender_accuracy mean: 0.924429 2020-02-01 21:21:43.876: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053761 2020-02-01 21:21:43.876: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862521 2020-02-01 21:21:43.876: INFO @log_variables: valid age_confidence/loss mean: 0.070289 2020-02-01 21:21:43.876: INFO @log_variables: valid age_confidence/accuracy mean: 0.565452 2020-02-01 21:21:43.877: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:21:43.884: INFO @metrics_hook: train age_mae: 5.648 +-0.033 (110592) 2020-02-01 21:21:43.891: INFO @metrics_hook: train gender_accuracy: 0.950 +-0.001 (110592) 2020-02-01 21:21:46.619: INFO @metrics_hook: valid age_mae: 6.305 +-0.091 (17639) 2020-02-01 21:21:46.620: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 21:21:48.089: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:21:48.089: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:21:48.089: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.20 2020-02-01 21:21:48.089: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:21:48.221: INFO @evaluate_confidence: Previous accuracy would be: 95.05 2020-02-01 21:21:48.222: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:21:48.283: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.89, 97.97, 98.03, 98.09, 98.17, 98.23, 98.3, 98.33, 98.38, 98.42, 98.47, 98.53, 98.57, 98.62, 98.67, 98.72, 98.77, 98.81, 98.86, 98.91, 98.96, 98.99, 99.04, 99.08, 99.1, 99.15, 99.18, 99.21, 99.26, 99.28, 99.3, 99.32, 99.35, 99.37, 99.4, 99.42, 99.44, 99.46, 99.48, 99.51, 99.54, 99.56, 99.59, 99.61, 99.63, 99.66] 2020-02-01 21:21:48.283: INFO @evaluate_confidence: Dropped ratios are: [11.42, 11.89, 12.35, 12.79, 13.26, 13.69, 14.13, 14.54, 14.98, 15.38, 15.8, 16.27, 16.74, 17.18, 17.63, 18.06, 18.5, 18.96, 19.39, 19.88, 20.34, 20.79, 21.27, 21.73, 22.2, 22.68, 23.16, 23.64, 24.19, 24.66, 25.2, 25.75, 26.25, 26.81, 27.36, 27.97, 28.53, 29.12, 29.73, 30.36, 31.0, 31.67, 32.38, 33.1, 33.86, 34.66] 2020-02-01 21:21:48.330: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:21:48.330: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:21:48.331: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:21:48.331: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 21:21:48.469: INFO @evaluate_confidence: Previous accuracy would be: 57.70 2020-02-01 21:21:48.470: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:21:48.487: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.57, 66.14, 66.8, 67.55, 68.32, 69.06, 69.85, 70.68] 2020-02-01 21:21:48.487: INFO @evaluate_confidence: Dropped ratios are: [40.07, 43.09, 46.19, 49.22, 52.15, 55.16, 57.94, 60.67] 2020-02-01 21:21:48.494: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:21:48.495: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.23 2020-02-01 21:21:48.495: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.23 2020-02-01 21:21:48.495: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 21:21:48.600: INFO @evaluate_confidence: Previous accuracy would be: 92.44 2020-02-01 21:21:48.600: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 21:21:48.609: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.41, 96.55, 96.66, 96.79, 96.89, 97.0, 97.07, 97.11, 97.19, 97.24, 97.32, 97.42, 97.51, 97.6, 97.66, 97.72, 97.8, 97.86, 97.91, 97.99, 98.03, 98.11, 98.19, 98.26, 98.28, 98.35, 98.42, 98.47, 98.54, 98.59, 98.65, 98.67, 98.69, 98.69, 98.71, 98.74, 98.8, 98.85, 98.93, 99.0, 99.04] 2020-02-01 21:21:48.609: INFO @evaluate_confidence: Dropped ratios are: [13.44, 13.86, 14.29, 14.74, 15.2, 15.64, 16.06, 16.47, 16.96, 17.31, 17.73, 18.25, 18.61, 19.11, 19.45, 19.85, 20.36, 20.93, 21.31, 21.77, 22.25, 22.65, 23.08, 23.58, 24.11, 24.59, 25.16, 25.74, 26.21, 26.86, 27.5, 28.03, 28.56, 29.12, 29.82, 30.53, 31.24, 31.96, 32.61, 33.44, 34.24] 2020-02-01 21:21:48.617: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:21:48.617: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 21:21:48.617: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 21:21:48.617: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.11 2020-02-01 21:21:48.746: INFO @evaluate_confidence: Previous accuracy would be: 53.12 2020-02-01 21:21:48.747: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 21:21:48.748: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.75, 59.77, 60.45, 60.58] 2020-02-01 21:21:48.748: INFO @evaluate_confidence: Dropped ratios are: [46.29, 50.59, 54.95, 59.16] 2020-02-01 21:21:48.800: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:21:49.524: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:21:49.608: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:21:50.062: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:21:50.316: INFO @decay_lr : LR updated to `5.6471756e-05` 2020-02-01 21:21:50.317: INFO @log_profile : T train: 121.811154 2020-02-01 21:21:50.317: INFO @log_profile : T valid: 5.495474 2020-02-01 21:21:50.318: INFO @log_profile : T read data: 1.965001 2020-02-01 21:21:50.318: INFO @log_profile : T hooks: 7.062767 2020-02-01 21:21:50.318: INFO @main_loop : Epoch 114 done 2020-02-01 21:21:50.318: INFO @main_loop : Training epoch 115 2020-02-01 21:24:01.004: INFO @log_variables: train loss nanmean: 0.710900 2020-02-01 21:24:01.004: INFO @log_variables: train age_loss mean: 5.175967 2020-02-01 21:24:01.004: INFO @log_variables: train gender_loss mean: 0.127479 2020-02-01 21:24:01.004: INFO @log_variables: train age_mae mean: 5.652213 2020-02-01 21:24:01.004: INFO @log_variables: train gender_accuracy mean: 0.949870 2020-02-01 21:24:01.004: INFO @log_variables: train gender_confidence/loss nanmean: 0.054241 2020-02-01 21:24:01.005: INFO @log_variables: train gender_confidence/accuracy mean: 0.847416 2020-02-01 21:24:01.005: INFO @log_variables: train age_confidence/loss mean: 0.070227 2020-02-01 21:24:01.005: INFO @log_variables: train age_confidence/accuracy mean: 0.609330 2020-02-01 21:24:01.005: INFO @log_variables: valid loss nanmean: 0.839579 2020-02-01 21:24:01.005: INFO @log_variables: valid age_loss mean: 5.860176 2020-02-01 21:24:01.005: INFO @log_variables: valid gender_loss mean: 0.200994 2020-02-01 21:24:01.005: INFO @log_variables: valid age_mae mean: 6.340155 2020-02-01 21:24:01.005: INFO @log_variables: valid gender_accuracy mean: 0.920404 2020-02-01 21:24:01.005: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053941 2020-02-01 21:24:01.005: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862180 2020-02-01 21:24:01.005: INFO @log_variables: valid age_confidence/loss mean: 0.070173 2020-02-01 21:24:01.005: INFO @log_variables: valid age_confidence/accuracy mean: 0.570100 2020-02-01 21:24:01.005: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:24:01.012: INFO @metrics_hook: train age_mae: 5.652 +-0.033 (110372) 2020-02-01 21:24:01.020: INFO @metrics_hook: train gender_accuracy: 0.950 +-0.001 (110372) 2020-02-01 21:24:03.787: INFO @metrics_hook: valid age_mae: 6.340 +-0.092 (17639) 2020-02-01 21:24:03.788: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 21:24:05.457: INFO @decay_lr : LR updated to `5.6189398e-05` 2020-02-01 21:24:05.459: INFO @log_profile : T train: 121.681480 2020-02-01 21:24:05.459: INFO @log_profile : T valid: 5.486238 2020-02-01 21:24:05.459: INFO @log_profile : T read data: 2.828070 2020-02-01 21:24:05.459: INFO @log_profile : T hooks: 5.069855 2020-02-01 21:24:05.459: INFO @main_loop : Epoch 115 done 2020-02-01 21:24:05.459: INFO @main_loop : Training epoch 116 2020-02-01 21:26:15.918: INFO @log_variables: train loss nanmean: 0.707731 2020-02-01 21:26:15.918: INFO @log_variables: train age_loss mean: 5.164187 2020-02-01 21:26:15.918: INFO @log_variables: train gender_loss mean: 0.126151 2020-02-01 21:26:15.918: INFO @log_variables: train age_mae mean: 5.640117 2020-02-01 21:26:15.918: INFO @log_variables: train gender_accuracy mean: 0.949888 2020-02-01 21:26:15.918: INFO @log_variables: train gender_confidence/loss nanmean: 0.053345 2020-02-01 21:26:15.918: INFO @log_variables: train gender_confidence/accuracy mean: 0.849482 2020-02-01 21:26:15.918: INFO @log_variables: train age_confidence/loss mean: 0.070232 2020-02-01 21:26:15.918: INFO @log_variables: train age_confidence/accuracy mean: 0.608379 2020-02-01 21:26:15.918: INFO @log_variables: valid loss nanmean: 0.837105 2020-02-01 21:26:15.918: INFO @log_variables: valid age_loss mean: 5.798183 2020-02-01 21:26:15.918: INFO @log_variables: valid gender_loss mean: 0.205044 2020-02-01 21:26:15.918: INFO @log_variables: valid age_mae mean: 6.278458 2020-02-01 21:26:15.919: INFO @log_variables: valid gender_accuracy mean: 0.921764 2020-02-01 21:26:15.919: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053297 2020-02-01 21:26:15.919: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874709 2020-02-01 21:26:15.919: INFO @log_variables: valid age_confidence/loss mean: 0.070297 2020-02-01 21:26:15.919: INFO @log_variables: valid age_confidence/accuracy mean: 0.564658 2020-02-01 21:26:15.919: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:26:15.926: INFO @metrics_hook: train age_mae: 5.640 +-0.032 (110372) 2020-02-01 21:26:15.933: INFO @metrics_hook: train gender_accuracy: 0.950 +-0.001 (110372) 2020-02-01 21:26:18.662: INFO @metrics_hook: valid age_mae: 6.278 +-0.089 (17639) 2020-02-01 21:26:18.663: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 21:26:20.125: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:26:20.126: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:26:20.126: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.36 +- 0.21 2020-02-01 21:26:20.126: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:26:20.259: INFO @evaluate_confidence: Previous accuracy would be: 94.99 2020-02-01 21:26:20.260: INFO @evaluate_confidence: Possible optimal thresholds are: [0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:26:20.322: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.85, 97.93, 97.99, 98.07, 98.14, 98.2, 98.26, 98.32, 98.37, 98.43, 98.5, 98.55, 98.6, 98.64, 98.68, 98.73, 98.77, 98.82, 98.87, 98.92, 98.95, 98.99, 99.02, 99.05, 99.07, 99.11, 99.13, 99.16, 99.2, 99.23, 99.27, 99.29, 99.32, 99.34, 99.37, 99.41, 99.43, 99.46, 99.49, 99.51, 99.54, 99.56, 99.57, 99.6, 99.62, 99.63] 2020-02-01 21:26:20.322: INFO @evaluate_confidence: Dropped ratios are: [11.62, 12.08, 12.54, 12.98, 13.44, 13.87, 14.31, 14.76, 15.22, 15.7, 16.13, 16.56, 17.01, 17.47, 17.9, 18.34, 18.78, 19.23, 19.69, 20.15, 20.62, 21.07, 21.55, 22.03, 22.54, 23.01, 23.47, 23.97, 24.46, 24.98, 25.51, 26.04, 26.55, 27.1, 27.69, 28.27, 28.85, 29.43, 30.08, 30.75, 31.39, 32.06, 32.79, 33.46, 34.21, 35.05] 2020-02-01 21:26:20.373: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:26:20.373: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:26:20.373: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.12 2020-02-01 21:26:20.374: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 21:26:20.512: INFO @evaluate_confidence: Previous accuracy would be: 57.59 2020-02-01 21:26:20.512: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:26:20.529: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [65.59, 66.26, 66.95, 67.68, 68.38, 69.12, 69.95, 70.93] 2020-02-01 21:26:20.529: INFO @evaluate_confidence: Dropped ratios are: [40.23, 43.31, 46.42, 49.48, 52.38, 55.26, 58.12, 60.86] 2020-02-01 21:26:20.537: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:26:20.537: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 21:26:20.537: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.24 2020-02-01 21:26:20.537: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.24 2020-02-01 21:26:20.645: INFO @evaluate_confidence: Previous accuracy would be: 92.18 2020-02-01 21:26:20.646: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 21:26:20.655: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.21, 96.35, 96.47, 96.6, 96.7, 96.76, 96.84, 96.92, 97.0, 97.08, 97.12, 97.19, 97.29, 97.37, 97.45, 97.56, 97.61, 97.64, 97.69, 97.73, 97.75, 97.83, 97.88, 97.96, 98.04, 98.13, 98.15, 98.21, 98.24, 98.31, 98.37, 98.43, 98.46, 98.52, 98.59, 98.61, 98.68, 98.73, 98.8, 98.84, 98.94] 2020-02-01 21:26:20.655: INFO @evaluate_confidence: Dropped ratios are: [13.18, 13.61, 13.94, 14.33, 14.72, 15.19, 15.6, 15.98, 16.47, 16.92, 17.31, 17.81, 18.17, 18.72, 19.14, 19.66, 20.04, 20.46, 20.88, 21.24, 21.68, 22.13, 22.57, 23.04, 23.47, 24.03, 24.46, 24.93, 25.47, 26.03, 26.55, 27.12, 27.75, 28.42, 28.97, 29.62, 30.39, 31.14, 31.87, 32.82, 33.68] 2020-02-01 21:26:20.663: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:26:20.663: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.11 2020-02-01 21:26:20.663: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 21:26:20.663: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.11 2020-02-01 21:26:20.793: INFO @evaluate_confidence: Previous accuracy would be: 52.97 2020-02-01 21:26:20.793: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 21:26:20.795: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.5, 58.97, 59.49, 59.76, 60.51] 2020-02-01 21:26:20.795: INFO @evaluate_confidence: Dropped ratios are: [44.45, 48.53, 52.64, 57.1, 61.21] 2020-02-01 21:26:20.846: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:26:21.540: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:26:21.625: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:26:22.091: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:26:22.344: INFO @decay_lr : LR updated to `5.590845e-05` 2020-02-01 21:26:22.345: INFO @log_profile : T train: 121.516617 2020-02-01 21:26:22.345: INFO @log_profile : T valid: 5.418564 2020-02-01 21:26:22.345: INFO @log_profile : T read data: 2.821544 2020-02-01 21:26:22.345: INFO @log_profile : T hooks: 7.051746 2020-02-01 21:26:22.345: INFO @main_loop : Epoch 116 done 2020-02-01 21:26:22.346: INFO @main_loop : Training epoch 117 2020-02-01 21:28:32.399: INFO @log_variables: train loss nanmean: 0.707347 2020-02-01 21:28:32.399: INFO @log_variables: train age_loss mean: 5.149306 2020-02-01 21:28:32.399: INFO @log_variables: train gender_loss mean: 0.126335 2020-02-01 21:28:32.399: INFO @log_variables: train age_mae mean: 5.625419 2020-02-01 21:28:32.399: INFO @log_variables: train gender_accuracy mean: 0.950638 2020-02-01 21:28:32.399: INFO @log_variables: train gender_confidence/loss nanmean: 0.053957 2020-02-01 21:28:32.399: INFO @log_variables: train gender_confidence/accuracy mean: 0.848045 2020-02-01 21:28:32.399: INFO @log_variables: train age_confidence/loss mean: 0.070422 2020-02-01 21:28:32.399: INFO @log_variables: train age_confidence/accuracy mean: 0.608724 2020-02-01 21:28:32.400: INFO @log_variables: valid loss nanmean: 0.830232 2020-02-01 21:28:32.400: INFO @log_variables: valid age_loss mean: 5.766889 2020-02-01 21:28:32.400: INFO @log_variables: valid gender_loss mean: 0.200700 2020-02-01 21:28:32.400: INFO @log_variables: valid age_mae mean: 6.246876 2020-02-01 21:28:32.400: INFO @log_variables: valid gender_accuracy mean: 0.923522 2020-02-01 21:28:32.400: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053622 2020-02-01 21:28:32.400: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867283 2020-02-01 21:28:32.400: INFO @log_variables: valid age_confidence/loss mean: 0.069892 2020-02-01 21:28:32.400: INFO @log_variables: valid age_confidence/accuracy mean: 0.560179 2020-02-01 21:28:32.400: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:28:32.407: INFO @metrics_hook: train age_mae: 5.625 +-0.032 (110592) 2020-02-01 21:28:32.414: INFO @metrics_hook: train gender_accuracy: 0.951 +-0.001 (110592) 2020-02-01 21:28:35.115: INFO @metrics_hook: valid age_mae: 6.247 +-0.087 (17639) 2020-02-01 21:28:35.116: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 21:28:36.760: INFO @decay_lr : LR updated to `5.562891e-05` 2020-02-01 21:28:36.762: INFO @log_profile : T train: 121.918328 2020-02-01 21:28:36.762: INFO @log_profile : T valid: 5.540744 2020-02-01 21:28:36.762: INFO @log_profile : T read data: 1.895073 2020-02-01 21:28:36.762: INFO @log_profile : T hooks: 4.985056 2020-02-01 21:28:36.762: INFO @main_loop : Epoch 117 done 2020-02-01 21:28:36.762: INFO @main_loop : Training epoch 118 2020-02-01 21:30:47.263: INFO @log_variables: train loss nanmean: 0.707022 2020-02-01 21:30:47.263: INFO @log_variables: train age_loss mean: 5.167732 2020-02-01 21:30:47.263: INFO @log_variables: train gender_loss mean: 0.125419 2020-02-01 21:30:47.263: INFO @log_variables: train age_mae mean: 5.643647 2020-02-01 21:30:47.263: INFO @log_variables: train gender_accuracy mean: 0.950051 2020-02-01 21:30:47.263: INFO @log_variables: train gender_confidence/loss nanmean: 0.052824 2020-02-01 21:30:47.263: INFO @log_variables: train gender_confidence/accuracy mean: 0.849137 2020-02-01 21:30:47.263: INFO @log_variables: train age_confidence/loss mean: 0.070387 2020-02-01 21:30:47.263: INFO @log_variables: train age_confidence/accuracy mean: 0.608062 2020-02-01 21:30:47.263: INFO @log_variables: valid loss nanmean: 0.828014 2020-02-01 21:30:47.263: INFO @log_variables: valid age_loss mean: 5.803004 2020-02-01 21:30:47.263: INFO @log_variables: valid gender_loss mean: 0.195387 2020-02-01 21:30:47.264: INFO @log_variables: valid age_mae mean: 6.283015 2020-02-01 21:30:47.264: INFO @log_variables: valid gender_accuracy mean: 0.924769 2020-02-01 21:30:47.264: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052698 2020-02-01 21:30:47.264: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873292 2020-02-01 21:30:47.264: INFO @log_variables: valid age_confidence/loss mean: 0.070145 2020-02-01 21:30:47.264: INFO @log_variables: valid age_confidence/accuracy mean: 0.557571 2020-02-01 21:30:47.264: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:30:47.271: INFO @metrics_hook: train age_mae: 5.644 +-0.033 (110372) 2020-02-01 21:30:47.279: INFO @metrics_hook: train gender_accuracy: 0.950 +-0.001 (110372) 2020-02-01 21:30:50.024: INFO @metrics_hook: valid age_mae: 6.283 +-0.089 (17639) 2020-02-01 21:30:50.025: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 21:30:51.495: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:30:51.496: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:30:51.496: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.20 2020-02-01 21:30:51.496: INFO @evaluate_confidence: Average confidence of all samples 0.79 +- 0.26 2020-02-01 21:30:51.625: INFO @evaluate_confidence: Previous accuracy would be: 95.01 2020-02-01 21:30:51.625: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:30:51.686: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.82, 97.91, 97.99, 98.07, 98.15, 98.22, 98.29, 98.36, 98.41, 98.47, 98.52, 98.58, 98.62, 98.67, 98.72, 98.77, 98.8, 98.85, 98.89, 98.93, 98.96, 99.0, 99.04, 99.06, 99.1, 99.14, 99.17, 99.21, 99.24, 99.25, 99.29, 99.33, 99.35, 99.37, 99.39, 99.41, 99.43, 99.45, 99.48, 99.5, 99.52, 99.54, 99.56, 99.58, 99.6, 99.61, 99.63] 2020-02-01 21:30:51.686: INFO @evaluate_confidence: Dropped ratios are: [11.4, 11.84, 12.3, 12.76, 13.23, 13.66, 14.12, 14.56, 15.01, 15.44, 15.88, 16.33, 16.76, 17.18, 17.63, 18.07, 18.49, 18.92, 19.36, 19.79, 20.26, 20.7, 21.16, 21.62, 22.04, 22.5, 22.97, 23.46, 23.91, 24.42, 24.9, 25.39, 25.89, 26.42, 26.94, 27.51, 28.09, 28.7, 29.28, 29.87, 30.5, 31.16, 31.84, 32.55, 33.29, 34.07, 34.85] 2020-02-01 21:30:51.735: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:30:51.735: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:30:51.735: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:30:51.736: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:30:51.872: INFO @evaluate_confidence: Previous accuracy would be: 57.71 2020-02-01 21:30:51.872: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:30:51.887: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.2, 66.86, 67.59, 68.32, 69.23, 69.96, 70.8] 2020-02-01 21:30:51.887: INFO @evaluate_confidence: Dropped ratios are: [42.84, 46.04, 49.25, 52.3, 55.16, 57.99, 60.73] 2020-02-01 21:30:51.894: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:30:51.894: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:30:51.894: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.23 2020-02-01 21:30:51.895: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 21:30:51.995: INFO @evaluate_confidence: Previous accuracy would be: 92.48 2020-02-01 21:30:51.995: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:30:52.004: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.4, 96.53, 96.66, 96.75, 96.83, 96.88, 96.98, 97.1, 97.2, 97.26, 97.36, 97.44, 97.5, 97.57, 97.66, 97.72, 97.75, 97.85, 97.91, 98.0, 98.03, 98.08, 98.12, 98.19, 98.22, 98.24, 98.28, 98.33, 98.37, 98.43, 98.47, 98.52, 98.54, 98.58, 98.66, 98.7, 98.75, 98.81, 98.85, 98.91, 98.92, 98.94] 2020-02-01 21:30:52.004: INFO @evaluate_confidence: Dropped ratios are: [12.69, 13.1, 13.55, 13.9, 14.28, 14.63, 15.07, 15.53, 15.91, 16.34, 16.8, 17.2, 17.58, 18.05, 18.5, 19.02, 19.48, 19.93, 20.36, 20.82, 21.32, 21.81, 22.25, 22.73, 23.27, 23.86, 24.33, 24.89, 25.47, 25.98, 26.5, 27.11, 27.7, 28.45, 29.17, 29.75, 30.53, 31.22, 32.1, 32.97, 33.75, 34.51] 2020-02-01 21:30:52.012: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:30:52.012: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:30:52.012: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 21:30:52.013: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.11 2020-02-01 21:30:52.137: INFO @evaluate_confidence: Previous accuracy would be: 52.62 2020-02-01 21:30:52.138: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51] 2020-02-01 21:30:52.139: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.42, 58.71, 59.24] 2020-02-01 21:30:52.139: INFO @evaluate_confidence: Dropped ratios are: [47.37, 51.89, 56.39] 2020-02-01 21:30:52.192: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:30:52.881: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:30:52.965: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:30:53.421: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:30:53.669: INFO @decay_lr : LR updated to `5.5350763e-05` 2020-02-01 21:30:53.670: INFO @log_profile : T train: 121.468296 2020-02-01 21:30:53.670: INFO @log_profile : T valid: 5.438669 2020-02-01 21:30:53.670: INFO @log_profile : T read data: 2.894153 2020-02-01 21:30:53.670: INFO @log_profile : T hooks: 7.028689 2020-02-01 21:30:53.670: INFO @main_loop : Epoch 118 done 2020-02-01 21:30:53.670: INFO @main_loop : Training epoch 119 2020-02-01 21:33:04.473: INFO @log_variables: train loss nanmean: 0.703225 2020-02-01 21:33:04.473: INFO @log_variables: train age_loss mean: 5.121903 2020-02-01 21:33:04.473: INFO @log_variables: train gender_loss mean: 0.124829 2020-02-01 21:33:04.473: INFO @log_variables: train age_mae mean: 5.597125 2020-02-01 21:33:04.473: INFO @log_variables: train gender_accuracy mean: 0.951138 2020-02-01 21:33:04.473: INFO @log_variables: train gender_confidence/loss nanmean: 0.053576 2020-02-01 21:33:04.473: INFO @log_variables: train gender_confidence/accuracy mean: 0.851412 2020-02-01 21:33:04.473: INFO @log_variables: train age_confidence/loss mean: 0.070541 2020-02-01 21:33:04.473: INFO @log_variables: train age_confidence/accuracy mean: 0.609774 2020-02-01 21:33:04.473: INFO @log_variables: valid loss nanmean: 0.841343 2020-02-01 21:33:04.474: INFO @log_variables: valid age_loss mean: 5.858627 2020-02-01 21:33:04.474: INFO @log_variables: valid gender_loss mean: 0.204332 2020-02-01 21:33:04.474: INFO @log_variables: valid age_mae mean: 6.338414 2020-02-01 21:33:04.474: INFO @log_variables: valid gender_accuracy mean: 0.921141 2020-02-01 21:33:04.474: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053431 2020-02-01 21:33:04.474: INFO @log_variables: valid gender_confidence/accuracy mean: 0.880946 2020-02-01 21:33:04.474: INFO @log_variables: valid age_confidence/loss mean: 0.069553 2020-02-01 21:33:04.474: INFO @log_variables: valid age_confidence/accuracy mean: 0.559272 2020-02-01 21:33:04.474: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:33:04.481: INFO @metrics_hook: train age_mae: 5.597 +-0.032 (110372) 2020-02-01 21:33:04.488: INFO @metrics_hook: train gender_accuracy: 0.951 +-0.001 (110372) 2020-02-01 21:33:07.223: INFO @metrics_hook: valid age_mae: 6.338 +-0.089 (17639) 2020-02-01 21:33:07.224: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 21:33:08.861: INFO @decay_lr : LR updated to `5.5074008e-05` 2020-02-01 21:33:08.863: INFO @log_profile : T train: 121.700134 2020-02-01 21:33:08.863: INFO @log_profile : T valid: 5.482790 2020-02-01 21:33:08.863: INFO @log_profile : T read data: 2.923188 2020-02-01 21:33:08.863: INFO @log_profile : T hooks: 5.009440 2020-02-01 21:33:08.863: INFO @main_loop : Epoch 119 done 2020-02-01 21:33:08.863: INFO @main_loop : Training epoch 120 2020-02-01 21:35:20.489: INFO @log_variables: train loss nanmean: 0.703771 2020-02-01 21:35:20.489: INFO @log_variables: train age_loss mean: 5.145679 2020-02-01 21:35:20.489: INFO @log_variables: train gender_loss mean: 0.123922 2020-02-01 21:35:20.489: INFO @log_variables: train age_mae mean: 5.621387 2020-02-01 21:35:20.489: INFO @log_variables: train gender_accuracy mean: 0.951136 2020-02-01 21:35:20.489: INFO @log_variables: train gender_confidence/loss nanmean: 0.052739 2020-02-01 21:35:20.489: INFO @log_variables: train gender_confidence/accuracy mean: 0.853425 2020-02-01 21:35:20.489: INFO @log_variables: train age_confidence/loss mean: 0.070586 2020-02-01 21:35:20.489: INFO @log_variables: train age_confidence/accuracy mean: 0.605415 2020-02-01 21:35:20.489: INFO @log_variables: valid loss nanmean: 0.834908 2020-02-01 21:35:20.489: INFO @log_variables: valid age_loss mean: 5.823596 2020-02-01 21:35:20.489: INFO @log_variables: valid gender_loss mean: 0.201132 2020-02-01 21:35:20.490: INFO @log_variables: valid age_mae mean: 6.304188 2020-02-01 21:35:20.490: INFO @log_variables: valid gender_accuracy mean: 0.922501 2020-02-01 21:35:20.490: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052604 2020-02-01 21:35:20.490: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874539 2020-02-01 21:35:20.490: INFO @log_variables: valid age_confidence/loss mean: 0.070039 2020-02-01 21:35:20.490: INFO @log_variables: valid age_confidence/accuracy mean: 0.556778 2020-02-01 21:35:20.490: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:35:20.497: INFO @metrics_hook: train age_mae: 5.621 +-0.032 (110592) 2020-02-01 21:35:20.504: INFO @metrics_hook: train gender_accuracy: 0.951 +-0.001 (110592) 2020-02-01 21:35:23.258: INFO @metrics_hook: valid age_mae: 6.304 +-0.089 (17639) 2020-02-01 21:35:23.259: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 21:35:24.727: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:35:24.728: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:35:24.728: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:35:24.728: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:35:24.856: INFO @evaluate_confidence: Previous accuracy would be: 95.11 2020-02-01 21:35:24.857: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:35:24.918: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.87, 97.97, 98.04, 98.1, 98.16, 98.2, 98.27, 98.33, 98.4, 98.45, 98.5, 98.55, 98.58, 98.63, 98.68, 98.73, 98.78, 98.83, 98.87, 98.92, 98.97, 99.0, 99.04, 99.07, 99.1, 99.14, 99.16, 99.19, 99.22, 99.24, 99.26, 99.3, 99.33, 99.37, 99.39, 99.41, 99.43, 99.46, 99.49, 99.5, 99.52, 99.55, 99.57, 99.6, 99.6, 99.62, 99.64] 2020-02-01 21:35:24.918: INFO @evaluate_confidence: Dropped ratios are: [10.85, 11.31, 11.78, 12.22, 12.63, 13.04, 13.48, 13.92, 14.36, 14.8, 15.23, 15.65, 16.1, 16.56, 17.0, 17.45, 17.9, 18.37, 18.81, 19.25, 19.7, 20.19, 20.69, 21.14, 21.58, 22.06, 22.54, 23.0, 23.53, 23.98, 24.47, 24.96, 25.49, 26.02, 26.58, 27.15, 27.71, 28.34, 28.94, 29.53, 30.16, 30.79, 31.46, 32.15, 32.88, 33.62, 34.38] 2020-02-01 21:35:24.969: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:35:24.969: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:35:24.969: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:35:24.969: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:35:25.116: INFO @evaluate_confidence: Previous accuracy would be: 57.95 2020-02-01 21:35:25.117: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:35:25.132: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.27, 66.89, 67.69, 68.43, 69.27, 69.99, 70.81] 2020-02-01 21:35:25.132: INFO @evaluate_confidence: Dropped ratios are: [43.17, 46.29, 49.48, 52.57, 55.68, 58.45, 61.18] 2020-02-01 21:35:25.140: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:35:25.140: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:35:25.140: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.24 2020-02-01 21:35:25.140: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 21:35:25.245: INFO @evaluate_confidence: Previous accuracy would be: 92.25 2020-02-01 21:35:25.246: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:35:25.254: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.39, 96.49, 96.59, 96.69, 96.76, 96.88, 96.95, 97.01, 97.1, 97.14, 97.19, 97.28, 97.33, 97.45, 97.47, 97.57, 97.63, 97.68, 97.76, 97.84, 97.87, 97.94, 98.0, 98.06, 98.1, 98.19, 98.23, 98.3, 98.39, 98.43, 98.47, 98.54, 98.54, 98.6, 98.65, 98.72, 98.76, 98.81, 98.85, 98.9, 98.94] 2020-02-01 21:35:25.254: INFO @evaluate_confidence: Dropped ratios are: [13.01, 13.35, 13.73, 14.11, 14.55, 14.99, 15.32, 15.81, 16.21, 16.68, 16.98, 17.49, 17.93, 18.37, 18.76, 19.19, 19.61, 20.07, 20.57, 21.11, 21.61, 22.08, 22.48, 22.91, 23.4, 24.0, 24.56, 25.1, 25.65, 26.17, 26.8, 27.46, 27.96, 28.49, 29.08, 29.87, 30.65, 31.33, 32.09, 32.89, 33.83] 2020-02-01 21:35:25.262: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:35:25.262: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 21:35:25.262: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.09 2020-02-01 21:35:25.263: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 21:35:25.389: INFO @evaluate_confidence: Previous accuracy would be: 52.37 2020-02-01 21:35:25.389: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 21:35:25.390: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.6, 57.88, 58.09, 58.61] 2020-02-01 21:35:25.390: INFO @evaluate_confidence: Dropped ratios are: [44.45, 48.94, 53.67, 58.18] 2020-02-01 21:35:25.441: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:35:26.140: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:35:26.227: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:35:26.683: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:35:26.759: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:35:27.445: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:35:27.529: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:35:27.531: INFO @evaluate_gender-age_model: groups 0 3.461923 1 4.099542 2 5.310349 3 5.620624 4 6.236231 5 6.267228 6 6.323977 7 7.001542 Name: errors, dtype: float64 2020-02-01 21:35:27.532: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:35:27.988: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:35:28.048: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:35:28.049: INFO @evaluate_gender-age_model: groups 0 6.312635 1 5.291444 2 5.606967 3 5.665922 4 7.379424 5 5.605290 6 7.657149 7 11.338937 Name: errors, dtype: float64 2020-02-01 21:35:28.230: INFO @decay_lr : LR updated to `5.479864e-05` 2020-02-01 21:35:28.231: INFO @log_profile : T train: 121.769039 2020-02-01 21:35:28.231: INFO @log_profile : T valid: 5.413326 2020-02-01 21:35:28.231: INFO @log_profile : T read data: 1.943871 2020-02-01 21:35:28.231: INFO @log_profile : T hooks: 10.167000 2020-02-01 21:35:28.232: INFO @main_loop : Epoch 120 done 2020-02-01 21:35:28.232: INFO @main_loop : Training epoch 121 2020-02-01 21:37:39.061: INFO @log_variables: train loss nanmean: 0.703339 2020-02-01 21:37:39.061: INFO @log_variables: train age_loss mean: 5.119836 2020-02-01 21:37:39.061: INFO @log_variables: train gender_loss mean: 0.125391 2020-02-01 21:37:39.061: INFO @log_variables: train age_mae mean: 5.595631 2020-02-01 21:37:39.061: INFO @log_variables: train gender_accuracy mean: 0.950821 2020-02-01 21:37:39.061: INFO @log_variables: train gender_confidence/loss nanmean: 0.053374 2020-02-01 21:37:39.061: INFO @log_variables: train gender_confidence/accuracy mean: 0.849808 2020-02-01 21:37:39.061: INFO @log_variables: train age_confidence/loss mean: 0.070534 2020-02-01 21:37:39.061: INFO @log_variables: train age_confidence/accuracy mean: 0.606947 2020-02-01 21:37:39.062: INFO @log_variables: valid loss nanmean: 0.820355 2020-02-01 21:37:39.062: INFO @log_variables: valid age_loss mean: 5.744421 2020-02-01 21:37:39.062: INFO @log_variables: valid gender_loss mean: 0.192160 2020-02-01 21:37:39.062: INFO @log_variables: valid age_mae mean: 6.224469 2020-02-01 21:37:39.062: INFO @log_variables: valid gender_accuracy mean: 0.926016 2020-02-01 21:37:39.062: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053162 2020-02-01 21:37:39.062: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876183 2020-02-01 21:37:39.062: INFO @log_variables: valid age_confidence/loss mean: 0.070281 2020-02-01 21:37:39.062: INFO @log_variables: valid age_confidence/accuracy mean: 0.561483 2020-02-01 21:37:39.062: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:37:39.069: INFO @metrics_hook: train age_mae: 5.596 +-0.032 (110372) 2020-02-01 21:37:39.076: INFO @metrics_hook: train gender_accuracy: 0.951 +-0.001 (110372) 2020-02-01 21:37:41.784: INFO @metrics_hook: valid age_mae: 6.224 +-0.088 (17639) 2020-02-01 21:37:41.786: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 21:37:43.425: INFO @decay_lr : LR updated to `5.4524644e-05` 2020-02-01 21:37:43.742: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 21:37:43.745: INFO @log_profile : T train: 121.801466 2020-02-01 21:37:43.745: INFO @log_profile : T valid: 5.443037 2020-02-01 21:37:43.745: INFO @log_profile : T read data: 2.887810 2020-02-01 21:37:43.745: INFO @log_profile : T hooks: 5.305262 2020-02-01 21:37:43.746: INFO @main_loop : Epoch 121 done 2020-02-01 21:37:43.746: INFO @main_loop : Training epoch 122 2020-02-01 21:39:54.325: INFO @log_variables: train loss nanmean: 0.700815 2020-02-01 21:39:54.325: INFO @log_variables: train age_loss mean: 5.135389 2020-02-01 21:39:54.325: INFO @log_variables: train gender_loss mean: 0.121301 2020-02-01 21:39:54.325: INFO @log_variables: train age_mae mean: 5.611217 2020-02-01 21:39:54.325: INFO @log_variables: train gender_accuracy mean: 0.952742 2020-02-01 21:39:54.325: INFO @log_variables: train gender_confidence/loss nanmean: 0.053238 2020-02-01 21:39:54.325: INFO @log_variables: train gender_confidence/accuracy mean: 0.851765 2020-02-01 21:39:54.325: INFO @log_variables: train age_confidence/loss mean: 0.070449 2020-02-01 21:39:54.325: INFO @log_variables: train age_confidence/accuracy mean: 0.606440 2020-02-01 21:39:54.325: INFO @log_variables: valid loss nanmean: 0.852690 2020-02-01 21:39:54.325: INFO @log_variables: valid age_loss mean: 5.775331 2020-02-01 21:39:54.325: INFO @log_variables: valid gender_loss mean: 0.221430 2020-02-01 21:39:54.325: INFO @log_variables: valid age_mae mean: 6.255241 2020-02-01 21:39:54.326: INFO @log_variables: valid gender_accuracy mean: 0.918249 2020-02-01 21:39:54.326: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055973 2020-02-01 21:39:54.326: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874256 2020-02-01 21:39:54.326: INFO @log_variables: valid age_confidence/loss mean: 0.070388 2020-02-01 21:39:54.326: INFO @log_variables: valid age_confidence/accuracy mean: 0.548160 2020-02-01 21:39:54.326: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:39:54.333: INFO @metrics_hook: train age_mae: 5.611 +-0.032 (110372) 2020-02-01 21:39:54.340: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 21:39:57.082: INFO @metrics_hook: valid age_mae: 6.255 +-0.088 (17639) 2020-02-01 21:39:57.084: INFO @metrics_hook: valid gender_accuracy: 0.918 +-0.004 (17639) 2020-02-01 21:39:58.558: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:39:58.558: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:39:58.558: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:39:58.558: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:39:58.690: INFO @evaluate_confidence: Previous accuracy would be: 95.27 2020-02-01 21:39:58.691: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:39:58.752: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.94, 98.02, 98.1, 98.15, 98.22, 98.27, 98.34, 98.39, 98.46, 98.5, 98.55, 98.6, 98.65, 98.69, 98.74, 98.79, 98.84, 98.88, 98.93, 98.96, 99.01, 99.03, 99.05, 99.08, 99.11, 99.13, 99.16, 99.19, 99.21, 99.25, 99.29, 99.3, 99.33, 99.35, 99.37, 99.4, 99.43, 99.45, 99.47, 99.49, 99.51, 99.53, 99.54, 99.57, 99.59, 99.61, 99.63] 2020-02-01 21:39:58.752: INFO @evaluate_confidence: Dropped ratios are: [10.98, 11.42, 11.89, 12.31, 12.8, 13.25, 13.69, 14.13, 14.59, 14.99, 15.44, 15.89, 16.28, 16.71, 17.12, 17.55, 17.99, 18.43, 18.86, 19.32, 19.77, 20.21, 20.65, 21.1, 21.57, 21.99, 22.42, 22.92, 23.36, 23.84, 24.33, 24.86, 25.34, 25.83, 26.33, 26.91, 27.46, 28.0, 28.57, 29.15, 29.76, 30.42, 31.09, 31.77, 32.53, 33.23, 34.04] 2020-02-01 21:39:58.801: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:39:58.801: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:39:58.801: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:39:58.802: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:39:58.939: INFO @evaluate_confidence: Previous accuracy would be: 57.79 2020-02-01 21:39:58.939: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:39:58.954: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.19, 66.98, 67.67, 68.3, 69.13, 70.01, 70.79] 2020-02-01 21:39:58.954: INFO @evaluate_confidence: Dropped ratios are: [43.05, 46.08, 49.22, 52.25, 55.19, 58.03, 60.72] 2020-02-01 21:39:58.962: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:39:58.962: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:39:58.962: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.24 2020-02-01 21:39:58.962: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 21:39:59.068: INFO @evaluate_confidence: Previous accuracy would be: 91.82 2020-02-01 21:39:59.068: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:39:59.077: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.14, 96.2, 96.34, 96.43, 96.54, 96.62, 96.69, 96.8, 96.85, 96.92, 96.95, 97.0, 97.11, 97.23, 97.29, 97.35, 97.38, 97.46, 97.55, 97.64, 97.71, 97.75, 97.8, 97.86, 97.92, 97.97, 98.03, 98.11, 98.16, 98.19, 98.25, 98.31, 98.35, 98.44, 98.44, 98.47, 98.58, 98.62, 98.69] 2020-02-01 21:39:59.077: INFO @evaluate_confidence: Dropped ratios are: [13.37, 13.82, 14.28, 14.65, 15.06, 15.55, 16.0, 16.37, 16.83, 17.21, 17.56, 18.04, 18.57, 19.07, 19.55, 20.04, 20.45, 20.88, 21.44, 21.92, 22.51, 23.0, 23.52, 24.03, 24.56, 25.08, 25.63, 26.18, 26.76, 27.43, 28.12, 28.78, 29.42, 30.2, 30.85, 31.47, 32.24, 33.02, 33.9] 2020-02-01 21:39:59.085: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:39:59.085: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:39:59.085: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 21:39:59.085: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 21:39:59.211: INFO @evaluate_confidence: Previous accuracy would be: 52.60 2020-02-01 21:39:59.211: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51] 2020-02-01 21:39:59.212: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.89, 58.07, 58.23] 2020-02-01 21:39:59.212: INFO @evaluate_confidence: Dropped ratios are: [49.25, 54.07, 58.99] 2020-02-01 21:39:59.265: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:39:59.961: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:40:00.043: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:40:00.522: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:40:00.775: INFO @decay_lr : LR updated to `5.425202e-05` 2020-02-01 21:40:00.777: INFO @log_profile : T train: 121.604886 2020-02-01 21:40:00.777: INFO @log_profile : T valid: 5.455736 2020-02-01 21:40:00.777: INFO @log_profile : T read data: 2.809821 2020-02-01 21:40:00.777: INFO @log_profile : T hooks: 7.082578 2020-02-01 21:40:00.777: INFO @main_loop : Epoch 122 done 2020-02-01 21:40:00.777: INFO @main_loop : Training epoch 123 2020-02-01 21:42:11.047: INFO @log_variables: train loss nanmean: 0.701712 2020-02-01 21:42:11.047: INFO @log_variables: train age_loss mean: 5.135739 2020-02-01 21:42:11.047: INFO @log_variables: train gender_loss mean: 0.122066 2020-02-01 21:42:11.048: INFO @log_variables: train age_mae mean: 5.611365 2020-02-01 21:42:11.048: INFO @log_variables: train gender_accuracy mean: 0.952610 2020-02-01 21:42:11.048: INFO @log_variables: train gender_confidence/loss nanmean: 0.053360 2020-02-01 21:42:11.048: INFO @log_variables: train gender_confidence/accuracy mean: 0.852747 2020-02-01 21:42:11.048: INFO @log_variables: train age_confidence/loss mean: 0.070498 2020-02-01 21:42:11.048: INFO @log_variables: train age_confidence/accuracy mean: 0.611021 2020-02-01 21:42:11.048: INFO @log_variables: valid loss nanmean: 0.841728 2020-02-01 21:42:11.048: INFO @log_variables: valid age_loss mean: 5.803881 2020-02-01 21:42:11.048: INFO @log_variables: valid gender_loss mean: 0.206111 2020-02-01 21:42:11.048: INFO @log_variables: valid age_mae mean: 6.284094 2020-02-01 21:42:11.048: INFO @log_variables: valid gender_accuracy mean: 0.922841 2020-02-01 21:42:11.048: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056251 2020-02-01 21:42:11.048: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870741 2020-02-01 21:42:11.048: INFO @log_variables: valid age_confidence/loss mean: 0.070477 2020-02-01 21:42:11.048: INFO @log_variables: valid age_confidence/accuracy mean: 0.558478 2020-02-01 21:42:11.048: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:42:11.056: INFO @metrics_hook: train age_mae: 5.611 +-0.032 (110592) 2020-02-01 21:42:11.063: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110592) 2020-02-01 21:42:13.823: INFO @metrics_hook: valid age_mae: 6.284 +-0.089 (17639) 2020-02-01 21:42:13.824: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 21:42:15.482: INFO @decay_lr : LR updated to `5.3980762e-05` 2020-02-01 21:42:15.484: INFO @log_profile : T train: 121.988203 2020-02-01 21:42:15.484: INFO @log_profile : T valid: 5.660801 2020-02-01 21:42:15.484: INFO @log_profile : T read data: 1.911384 2020-02-01 21:42:15.484: INFO @log_profile : T hooks: 5.068939 2020-02-01 21:42:15.484: INFO @main_loop : Epoch 123 done 2020-02-01 21:42:15.484: INFO @main_loop : Training epoch 124 2020-02-01 21:44:34.456: INFO @log_variables: train loss nanmean: 0.701089 2020-02-01 21:44:34.456: INFO @log_variables: train age_loss mean: 5.124932 2020-02-01 21:44:34.456: INFO @log_variables: train gender_loss mean: 0.122330 2020-02-01 21:44:34.456: INFO @log_variables: train age_mae mean: 5.600931 2020-02-01 21:44:34.456: INFO @log_variables: train gender_accuracy mean: 0.952669 2020-02-01 21:44:34.456: INFO @log_variables: train gender_confidence/loss nanmean: 0.053570 2020-02-01 21:44:34.456: INFO @log_variables: train gender_confidence/accuracy mean: 0.850578 2020-02-01 21:44:34.456: INFO @log_variables: train age_confidence/loss mean: 0.070407 2020-02-01 21:44:34.456: INFO @log_variables: train age_confidence/accuracy mean: 0.607926 2020-02-01 21:44:34.456: INFO @log_variables: valid loss nanmean: 0.838783 2020-02-01 21:44:34.457: INFO @log_variables: valid age_loss mean: 5.846444 2020-02-01 21:44:34.457: INFO @log_variables: valid gender_loss mean: 0.201786 2020-02-01 21:44:34.457: INFO @log_variables: valid age_mae mean: 6.326572 2020-02-01 21:44:34.457: INFO @log_variables: valid gender_accuracy mean: 0.922898 2020-02-01 21:44:34.457: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053872 2020-02-01 21:44:34.457: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868473 2020-02-01 21:44:34.457: INFO @log_variables: valid age_confidence/loss mean: 0.069975 2020-02-01 21:44:34.457: INFO @log_variables: valid age_confidence/accuracy mean: 0.553036 2020-02-01 21:44:34.457: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:44:34.464: INFO @metrics_hook: train age_mae: 5.601 +-0.032 (110372) 2020-02-01 21:44:34.471: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 21:44:37.191: INFO @metrics_hook: valid age_mae: 6.327 +-0.090 (17639) 2020-02-01 21:44:37.193: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 21:44:38.636: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:44:38.636: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:44:38.636: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:44:38.636: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:44:38.765: INFO @evaluate_confidence: Previous accuracy would be: 95.27 2020-02-01 21:44:38.766: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:44:38.825: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.92, 97.98, 98.07, 98.13, 98.21, 98.27, 98.31, 98.38, 98.44, 98.5, 98.54, 98.58, 98.63, 98.67, 98.71, 98.76, 98.8, 98.83, 98.88, 98.92, 98.96, 99.0, 99.03, 99.07, 99.09, 99.14, 99.16, 99.2, 99.23, 99.25, 99.27, 99.3, 99.33, 99.36, 99.39, 99.42, 99.43, 99.47, 99.49, 99.51, 99.53, 99.54, 99.56, 99.58, 99.6, 99.61, 99.63] 2020-02-01 21:44:38.825: INFO @evaluate_confidence: Dropped ratios are: [11.01, 11.46, 11.92, 12.37, 12.82, 13.28, 13.7, 14.12, 14.56, 15.0, 15.43, 15.81, 16.28, 16.73, 17.18, 17.63, 18.09, 18.5, 18.96, 19.42, 19.86, 20.31, 20.77, 21.22, 21.65, 22.14, 22.63, 23.1, 23.59, 24.09, 24.59, 25.07, 25.57, 26.07, 26.61, 27.19, 27.74, 28.32, 28.91, 29.53, 30.12, 30.76, 31.43, 32.13, 32.84, 33.58, 34.42] 2020-02-01 21:44:38.872: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:44:38.873: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:44:38.873: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:44:38.873: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:44:39.009: INFO @evaluate_confidence: Previous accuracy would be: 57.80 2020-02-01 21:44:39.009: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:44:39.024: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.33, 67.13, 67.73, 68.5, 69.26, 70.07, 70.88] 2020-02-01 21:44:39.024: INFO @evaluate_confidence: Dropped ratios are: [43.06, 46.22, 49.25, 52.27, 55.26, 58.14, 60.86] 2020-02-01 21:44:39.032: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:44:39.032: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:44:39.032: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.23 2020-02-01 21:44:39.032: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 21:44:39.135: INFO @evaluate_confidence: Previous accuracy would be: 92.29 2020-02-01 21:44:39.136: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:44:39.145: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.49, 96.61, 96.75, 96.81, 96.93, 96.99, 97.06, 97.14, 97.2, 97.25, 97.31, 97.38, 97.44, 97.48, 97.56, 97.63, 97.65, 97.74, 97.78, 97.86, 97.89, 97.93, 97.93, 97.94, 98.02, 98.08, 98.13, 98.17, 98.22, 98.29, 98.33, 98.43, 98.48, 98.54, 98.58, 98.64, 98.73, 98.78, 98.85, 98.87, 98.92] 2020-02-01 21:44:39.145: INFO @evaluate_confidence: Dropped ratios are: [13.48, 13.97, 14.5, 14.93, 15.39, 15.79, 16.29, 16.63, 17.04, 17.47, 17.94, 18.37, 18.84, 19.25, 19.62, 20.03, 20.44, 20.83, 21.26, 21.73, 22.27, 22.71, 23.2, 23.69, 24.19, 24.74, 25.28, 25.76, 26.33, 26.95, 27.5, 28.01, 28.63, 29.15, 29.76, 30.49, 31.24, 32.01, 32.83, 33.63, 34.46] 2020-02-01 21:44:39.152: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:44:39.153: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:44:39.153: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 21:44:39.153: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 21:44:39.280: INFO @evaluate_confidence: Previous accuracy would be: 52.42 2020-02-01 21:44:39.280: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 21:44:39.282: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.97, 58.03, 58.44, 58.92] 2020-02-01 21:44:39.282: INFO @evaluate_confidence: Dropped ratios are: [45.81, 49.95, 54.2, 58.46] 2020-02-01 21:44:39.340: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:44:40.039: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:44:40.126: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:44:40.586: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:44:40.850: INFO @decay_lr : LR updated to `5.3710857e-05` 2020-02-01 21:44:40.851: INFO @log_profile : T train: 130.014841 2020-02-01 21:44:40.851: INFO @log_profile : T valid: 5.368261 2020-02-01 21:44:40.851: INFO @log_profile : T read data: 2.910078 2020-02-01 21:44:40.851: INFO @log_profile : T hooks: 6.998184 2020-02-01 21:44:40.851: INFO @main_loop : Epoch 124 done 2020-02-01 21:44:40.852: INFO @main_loop : Training epoch 125 2020-02-01 21:46:51.790: INFO @log_variables: train loss nanmean: 0.697094 2020-02-01 21:46:51.790: INFO @log_variables: train age_loss mean: 5.123477 2020-02-01 21:46:51.790: INFO @log_variables: train gender_loss mean: 0.119830 2020-02-01 21:46:51.790: INFO @log_variables: train age_mae mean: 5.599037 2020-02-01 21:46:51.790: INFO @log_variables: train gender_accuracy mean: 0.952506 2020-02-01 21:46:51.790: INFO @log_variables: train gender_confidence/loss nanmean: 0.051837 2020-02-01 21:46:51.790: INFO @log_variables: train gender_confidence/accuracy mean: 0.855135 2020-02-01 21:46:51.790: INFO @log_variables: train age_confidence/loss mean: 0.070549 2020-02-01 21:46:51.791: INFO @log_variables: train age_confidence/accuracy mean: 0.606377 2020-02-01 21:46:51.791: INFO @log_variables: valid loss nanmean: 0.843275 2020-02-01 21:46:51.791: INFO @log_variables: valid age_loss mean: 5.828977 2020-02-01 21:46:51.791: INFO @log_variables: valid gender_loss mean: 0.206928 2020-02-01 21:46:51.791: INFO @log_variables: valid age_mae mean: 6.308187 2020-02-01 21:46:51.791: INFO @log_variables: valid gender_accuracy mean: 0.920857 2020-02-01 21:46:51.791: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054943 2020-02-01 21:46:51.791: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873576 2020-02-01 21:46:51.791: INFO @log_variables: valid age_confidence/loss mean: 0.070308 2020-02-01 21:46:51.791: INFO @log_variables: valid age_confidence/accuracy mean: 0.565055 2020-02-01 21:46:51.791: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:46:51.798: INFO @metrics_hook: train age_mae: 5.599 +-0.032 (110372) 2020-02-01 21:46:51.806: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 21:47:02.337: INFO @metrics_hook: valid age_mae: 6.308 +-0.090 (17639) 2020-02-01 21:47:02.339: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 21:47:04.020: INFO @decay_lr : LR updated to `5.34423e-05` 2020-02-01 21:47:04.021: INFO @log_profile : T train: 121.790015 2020-02-01 21:47:04.021: INFO @log_profile : T valid: 5.560424 2020-02-01 21:47:04.021: INFO @log_profile : T read data: 2.891056 2020-02-01 21:47:04.021: INFO @log_profile : T hooks: 12.850997 2020-02-01 21:47:04.021: INFO @main_loop : Epoch 125 done 2020-02-01 21:47:04.021: INFO @main_loop : Training epoch 126 2020-02-01 21:49:23.148: INFO @log_variables: train loss nanmean: 0.699728 2020-02-01 21:49:23.148: INFO @log_variables: train age_loss mean: 5.126044 2020-02-01 21:49:23.148: INFO @log_variables: train gender_loss mean: 0.121385 2020-02-01 21:49:23.148: INFO @log_variables: train age_mae mean: 5.601854 2020-02-01 21:49:23.148: INFO @log_variables: train gender_accuracy mean: 0.951841 2020-02-01 21:49:23.148: INFO @log_variables: train gender_confidence/loss nanmean: 0.052771 2020-02-01 21:49:23.148: INFO @log_variables: train gender_confidence/accuracy mean: 0.852512 2020-02-01 21:49:23.148: INFO @log_variables: train age_confidence/loss mean: 0.070603 2020-02-01 21:49:23.148: INFO @log_variables: train age_confidence/accuracy mean: 0.607232 2020-02-01 21:49:23.148: INFO @log_variables: valid loss nanmean: 0.830153 2020-02-01 21:49:23.148: INFO @log_variables: valid age_loss mean: 5.812480 2020-02-01 21:49:23.148: INFO @log_variables: valid gender_loss mean: 0.195083 2020-02-01 21:49:23.148: INFO @log_variables: valid age_mae mean: 6.292662 2020-02-01 21:49:23.148: INFO @log_variables: valid gender_accuracy mean: 0.925449 2020-02-01 21:49:23.148: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054576 2020-02-01 21:49:23.148: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873859 2020-02-01 21:49:23.149: INFO @log_variables: valid age_confidence/loss mean: 0.069821 2020-02-01 21:49:23.149: INFO @log_variables: valid age_confidence/accuracy mean: 0.565905 2020-02-01 21:49:23.149: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:49:23.156: INFO @metrics_hook: train age_mae: 5.602 +-0.032 (110592) 2020-02-01 21:49:23.163: INFO @metrics_hook: train gender_accuracy: 0.952 +-0.001 (110592) 2020-02-01 21:49:25.884: INFO @metrics_hook: valid age_mae: 6.293 +-0.089 (17639) 2020-02-01 21:49:25.885: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 21:49:27.338: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:49:27.338: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:49:27.339: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:49:27.339: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:49:27.469: INFO @evaluate_confidence: Previous accuracy would be: 95.18 2020-02-01 21:49:27.469: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:49:27.533: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.91, 98.0, 98.08, 98.16, 98.21, 98.27, 98.35, 98.41, 98.46, 98.51, 98.55, 98.6, 98.64, 98.69, 98.75, 98.79, 98.83, 98.86, 98.9, 98.94, 98.98, 99.02, 99.04, 99.08, 99.12, 99.14, 99.18, 99.2, 99.24, 99.27, 99.29, 99.32, 99.34, 99.36, 99.38, 99.4, 99.42, 99.44, 99.47, 99.49, 99.52, 99.54, 99.56, 99.58, 99.6, 99.61, 99.63] 2020-02-01 21:49:27.534: INFO @evaluate_confidence: Dropped ratios are: [11.15, 11.61, 12.05, 12.5, 12.9, 13.34, 13.78, 14.19, 14.61, 15.03, 15.46, 15.89, 16.29, 16.69, 17.14, 17.56, 17.99, 18.35, 18.82, 19.26, 19.73, 20.19, 20.63, 21.07, 21.49, 21.91, 22.37, 22.81, 23.3, 23.77, 24.26, 24.77, 25.28, 25.78, 26.33, 26.85, 27.38, 27.92, 28.5, 29.09, 29.7, 30.3, 30.97, 31.64, 32.3, 33.09, 33.9] 2020-02-01 21:49:27.585: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:49:27.585: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:49:27.585: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:49:27.585: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:49:27.720: INFO @evaluate_confidence: Previous accuracy would be: 57.97 2020-02-01 21:49:27.720: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:49:27.735: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.33, 66.97, 67.63, 68.41, 69.12, 69.91, 70.69] 2020-02-01 21:49:27.735: INFO @evaluate_confidence: Dropped ratios are: [42.78, 45.96, 49.06, 52.11, 54.96, 57.83, 60.57] 2020-02-01 21:49:27.742: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:49:27.742: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 21:49:27.743: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.24 2020-02-01 21:49:27.743: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 21:49:27.845: INFO @evaluate_confidence: Previous accuracy would be: 92.54 2020-02-01 21:49:27.846: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:49:27.854: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.52, 96.59, 96.61, 96.73, 96.79, 96.88, 96.97, 97.02, 97.09, 97.17, 97.29, 97.35, 97.4, 97.45, 97.48, 97.53, 97.56, 97.67, 97.77, 97.81, 97.91, 97.97, 98.04, 98.08, 98.11, 98.19, 98.21, 98.25, 98.29, 98.33, 98.43, 98.51, 98.6, 98.66, 98.72, 98.79, 98.83, 98.88, 98.9, 98.97, 99.0] 2020-02-01 21:49:27.854: INFO @evaluate_confidence: Dropped ratios are: [12.93, 13.29, 13.55, 13.87, 14.26, 14.75, 15.2, 15.53, 15.94, 16.37, 16.77, 17.22, 17.68, 18.08, 18.45, 18.72, 19.13, 19.56, 20.0, 20.36, 20.85, 21.28, 21.84, 22.33, 22.76, 23.31, 23.83, 24.29, 24.89, 25.56, 26.14, 26.75, 27.27, 27.9, 28.57, 29.33, 30.1, 30.79, 31.65, 32.43, 33.22] 2020-02-01 21:49:27.862: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:49:27.862: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.10 2020-02-01 21:49:27.862: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 21:49:27.862: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 21:49:27.987: INFO @evaluate_confidence: Previous accuracy would be: 52.43 2020-02-01 21:49:27.988: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 21:49:27.989: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.13, 58.23, 58.59, 58.95, 59.51] 2020-02-01 21:49:27.989: INFO @evaluate_confidence: Dropped ratios are: [44.4, 49.04, 53.55, 58.3, 62.98] 2020-02-01 21:49:28.042: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:49:28.740: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:49:28.829: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:49:29.280: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:49:29.533: INFO @decay_lr : LR updated to `5.317509e-05` 2020-02-01 21:49:29.534: INFO @log_profile : T train: 130.428839 2020-02-01 21:49:29.534: INFO @log_profile : T valid: 6.184371 2020-02-01 21:49:29.534: INFO @log_profile : T read data: 1.834245 2020-02-01 21:49:29.534: INFO @log_profile : T hooks: 6.989381 2020-02-01 21:49:29.534: INFO @main_loop : Epoch 126 done 2020-02-01 21:49:29.535: INFO @main_loop : Training epoch 127 2020-02-01 21:51:48.793: INFO @log_variables: train loss nanmean: 0.696510 2020-02-01 21:51:48.793: INFO @log_variables: train age_loss mean: 5.100613 2020-02-01 21:51:48.793: INFO @log_variables: train gender_loss mean: 0.120282 2020-02-01 21:51:48.793: INFO @log_variables: train age_mae mean: 5.575935 2020-02-01 21:51:48.793: INFO @log_variables: train gender_accuracy mean: 0.952887 2020-02-01 21:51:48.793: INFO @log_variables: train gender_confidence/loss nanmean: 0.053124 2020-02-01 21:51:48.793: INFO @log_variables: train gender_confidence/accuracy mean: 0.852852 2020-02-01 21:51:48.793: INFO @log_variables: train age_confidence/loss mean: 0.070347 2020-02-01 21:51:48.794: INFO @log_variables: train age_confidence/accuracy mean: 0.609820 2020-02-01 21:51:48.794: INFO @log_variables: valid loss nanmean: 0.824972 2020-02-01 21:51:48.794: INFO @log_variables: valid age_loss mean: 5.849797 2020-02-01 21:51:48.794: INFO @log_variables: valid gender_loss mean: 0.188001 2020-02-01 21:51:48.794: INFO @log_variables: valid age_mae mean: 6.330203 2020-02-01 21:51:48.794: INFO @log_variables: valid gender_accuracy mean: 0.927093 2020-02-01 21:51:48.794: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052885 2020-02-01 21:51:48.794: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866319 2020-02-01 21:51:48.794: INFO @log_variables: valid age_confidence/loss mean: 0.069378 2020-02-01 21:51:48.794: INFO @log_variables: valid age_confidence/accuracy mean: 0.566302 2020-02-01 21:51:48.794: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:51:48.801: INFO @metrics_hook: train age_mae: 5.576 +-0.032 (110372) 2020-02-01 21:51:48.808: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 21:51:51.509: INFO @metrics_hook: valid age_mae: 6.330 +-0.090 (17639) 2020-02-01 21:51:51.510: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 21:51:53.164: INFO @decay_lr : LR updated to `5.2909214e-05` 2020-02-01 21:51:53.493: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 21:51:53.496: INFO @log_profile : T train: 130.151507 2020-02-01 21:51:53.497: INFO @log_profile : T valid: 5.665006 2020-02-01 21:51:53.497: INFO @log_profile : T read data: 2.768904 2020-02-01 21:51:53.497: INFO @log_profile : T hooks: 5.301208 2020-02-01 21:51:53.497: INFO @main_loop : Epoch 127 done 2020-02-01 21:51:53.497: INFO @main_loop : Training epoch 128 2020-02-01 21:54:13.852: INFO @log_variables: train loss nanmean: 0.693391 2020-02-01 21:54:13.852: INFO @log_variables: train age_loss mean: 5.077388 2020-02-01 21:54:13.852: INFO @log_variables: train gender_loss mean: 0.119404 2020-02-01 21:54:13.853: INFO @log_variables: train age_mae mean: 5.553216 2020-02-01 21:54:13.853: INFO @log_variables: train gender_accuracy mean: 0.953349 2020-02-01 21:54:13.853: INFO @log_variables: train gender_confidence/loss nanmean: 0.052582 2020-02-01 21:54:13.853: INFO @log_variables: train gender_confidence/accuracy mean: 0.853903 2020-02-01 21:54:13.853: INFO @log_variables: train age_confidence/loss mean: 0.070679 2020-02-01 21:54:13.853: INFO @log_variables: train age_confidence/accuracy mean: 0.609856 2020-02-01 21:54:13.853: INFO @log_variables: valid loss nanmean: 0.834852 2020-02-01 21:54:13.853: INFO @log_variables: valid age_loss mean: 5.915065 2020-02-01 21:54:13.853: INFO @log_variables: valid gender_loss mean: 0.193905 2020-02-01 21:54:13.853: INFO @log_variables: valid age_mae mean: 6.395648 2020-02-01 21:54:13.853: INFO @log_variables: valid gender_accuracy mean: 0.925052 2020-02-01 21:54:13.853: INFO @log_variables: valid gender_confidence/loss nanmean: 0.051505 2020-02-01 21:54:13.853: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869437 2020-02-01 21:54:13.853: INFO @log_variables: valid age_confidence/loss mean: 0.069337 2020-02-01 21:54:13.853: INFO @log_variables: valid age_confidence/accuracy mean: 0.561767 2020-02-01 21:54:13.853: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:54:13.860: INFO @metrics_hook: train age_mae: 5.553 +-0.032 (110372) 2020-02-01 21:54:13.868: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 21:54:16.593: INFO @metrics_hook: valid age_mae: 6.396 +-0.091 (17639) 2020-02-01 21:54:16.594: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 21:54:18.020: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:54:18.021: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:54:18.021: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:54:18.021: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:54:18.148: INFO @evaluate_confidence: Previous accuracy would be: 95.33 2020-02-01 21:54:18.149: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:54:18.209: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [97.95, 98.02, 98.08, 98.15, 98.21, 98.27, 98.32, 98.38, 98.44, 98.5, 98.54, 98.58, 98.62, 98.67, 98.72, 98.77, 98.82, 98.86, 98.91, 98.96, 99.0, 99.04, 99.08, 99.11, 99.14, 99.18, 99.21, 99.24, 99.27, 99.29, 99.31, 99.34, 99.37, 99.39, 99.41, 99.43, 99.46, 99.48, 99.5, 99.52, 99.54, 99.57, 99.59, 99.61, 99.63, 99.64, 99.66] 2020-02-01 21:54:18.209: INFO @evaluate_confidence: Dropped ratios are: [10.84, 11.29, 11.73, 12.15, 12.56, 13.02, 13.43, 13.85, 14.27, 14.73, 15.15, 15.54, 15.93, 16.37, 16.79, 17.24, 17.68, 18.11, 18.57, 18.99, 19.44, 19.89, 20.32, 20.75, 21.19, 21.68, 22.14, 22.6, 23.09, 23.57, 24.04, 24.52, 25.02, 25.57, 26.12, 26.66, 27.2, 27.79, 28.39, 28.97, 29.58, 30.23, 30.91, 31.64, 32.29, 33.04, 33.77] 2020-02-01 21:54:18.258: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:54:18.259: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:54:18.259: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:54:18.259: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:54:18.396: INFO @evaluate_confidence: Previous accuracy would be: 58.16 2020-02-01 21:54:18.396: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:54:18.412: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.6, 67.22, 67.89, 68.55, 69.26, 69.98, 70.8] 2020-02-01 21:54:18.412: INFO @evaluate_confidence: Dropped ratios are: [42.32, 45.48, 48.58, 51.57, 54.52, 57.38, 60.14] 2020-02-01 21:54:18.420: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:54:18.420: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 21:54:18.420: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.42 +- 0.24 2020-02-01 21:54:18.420: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 21:54:18.527: INFO @evaluate_confidence: Previous accuracy would be: 92.51 2020-02-01 21:54:18.527: INFO @evaluate_confidence: Possible optimal thresholds are: [0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 21:54:18.537: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.64, 96.68, 96.74, 96.83, 96.9, 97.0, 97.1, 97.21, 97.29, 97.32, 97.43, 97.5, 97.55, 97.62, 97.7, 97.78, 97.81, 97.87, 97.91, 97.98, 98.03, 98.07, 98.11, 98.14, 98.2, 98.22, 98.28, 98.31, 98.34, 98.41, 98.45, 98.5, 98.54, 98.58, 98.61, 98.67, 98.71, 98.76, 98.79, 98.82, 98.86, 98.9, 98.93, 98.96] 2020-02-01 21:54:18.537: INFO @evaluate_confidence: Dropped ratios are: [12.99, 13.29, 13.72, 14.07, 14.41, 14.81, 15.24, 15.57, 15.99, 16.33, 16.75, 17.15, 17.53, 17.88, 18.29, 18.69, 19.2, 19.57, 19.94, 20.33, 20.76, 21.19, 21.61, 22.01, 22.48, 22.97, 23.44, 23.87, 24.3, 24.84, 25.4, 25.97, 26.54, 27.16, 27.56, 28.13, 28.69, 29.23, 29.84, 30.44, 31.16, 31.96, 32.8, 33.58] 2020-02-01 21:54:18.545: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:54:18.545: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 21:54:18.545: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.09 2020-02-01 21:54:18.545: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 21:54:18.673: INFO @evaluate_confidence: Previous accuracy would be: 51.81 2020-02-01 21:54:18.673: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 21:54:18.674: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.31, 57.71, 58.47, 59.52] 2020-02-01 21:54:18.674: INFO @evaluate_confidence: Dropped ratios are: [43.71, 48.18, 52.89, 57.67] 2020-02-01 21:54:18.726: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:54:19.411: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:54:19.498: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:54:19.957: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:54:20.219: INFO @decay_lr : LR updated to `5.2644667e-05` 2020-02-01 21:54:20.221: INFO @log_profile : T train: 130.063517 2020-02-01 21:54:20.221: INFO @log_profile : T valid: 6.780523 2020-02-01 21:54:20.221: INFO @log_profile : T read data: 2.838254 2020-02-01 21:54:20.221: INFO @log_profile : T hooks: 6.966108 2020-02-01 21:54:20.221: INFO @main_loop : Epoch 128 done 2020-02-01 21:54:20.221: INFO @main_loop : Training epoch 129 2020-02-01 21:56:33.947: INFO @log_variables: train loss nanmean: 0.694213 2020-02-01 21:56:33.948: INFO @log_variables: train age_loss mean: 5.081481 2020-02-01 21:56:33.948: INFO @log_variables: train gender_loss mean: 0.120244 2020-02-01 21:56:33.948: INFO @log_variables: train age_mae mean: 5.556656 2020-02-01 21:56:33.948: INFO @log_variables: train gender_accuracy mean: 0.952302 2020-02-01 21:56:33.948: INFO @log_variables: train gender_confidence/loss nanmean: 0.052347 2020-02-01 21:56:33.948: INFO @log_variables: train gender_confidence/accuracy mean: 0.852691 2020-02-01 21:56:33.948: INFO @log_variables: train age_confidence/loss mean: 0.070600 2020-02-01 21:56:33.948: INFO @log_variables: train age_confidence/accuracy mean: 0.608449 2020-02-01 21:56:33.948: INFO @log_variables: valid loss nanmean: 0.820486 2020-02-01 21:56:33.948: INFO @log_variables: valid age_loss mean: 5.702569 2020-02-01 21:56:33.948: INFO @log_variables: valid gender_loss mean: 0.196300 2020-02-01 21:56:33.948: INFO @log_variables: valid age_mae mean: 6.181912 2020-02-01 21:56:33.948: INFO @log_variables: valid gender_accuracy mean: 0.927377 2020-02-01 21:56:33.948: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053274 2020-02-01 21:56:33.948: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877828 2020-02-01 21:56:33.948: INFO @log_variables: valid age_confidence/loss mean: 0.070343 2020-02-01 21:56:33.948: INFO @log_variables: valid age_confidence/accuracy mean: 0.567549 2020-02-01 21:56:33.948: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:56:33.956: INFO @metrics_hook: train age_mae: 5.557 +-0.032 (110591) 2020-02-01 21:56:33.963: INFO @metrics_hook: train gender_accuracy: 0.952 +-0.001 (110591) 2020-02-01 21:56:36.732: INFO @metrics_hook: valid age_mae: 6.182 +-0.088 (17639) 2020-02-01 21:56:36.733: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 21:56:38.389: INFO @decay_lr : LR updated to `5.2381445e-05` 2020-02-01 21:56:38.709: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 21:56:38.713: INFO @log_profile : T train: 125.716210 2020-02-01 21:56:38.713: INFO @log_profile : T valid: 5.450812 2020-02-01 21:56:38.713: INFO @log_profile : T read data: 1.869378 2020-02-01 21:56:38.713: INFO @log_profile : T hooks: 5.379809 2020-02-01 21:56:38.713: INFO @main_loop : Epoch 129 done 2020-02-01 21:56:38.713: INFO @main_loop : Training epoch 130 2020-02-01 21:58:52.399: INFO @log_variables: train loss nanmean: 0.691479 2020-02-01 21:58:52.399: INFO @log_variables: train age_loss mean: 5.063249 2020-02-01 21:58:52.399: INFO @log_variables: train gender_loss mean: 0.118025 2020-02-01 21:58:52.399: INFO @log_variables: train age_mae mean: 5.539072 2020-02-01 21:58:52.399: INFO @log_variables: train gender_accuracy mean: 0.953747 2020-02-01 21:58:52.399: INFO @log_variables: train gender_confidence/loss nanmean: 0.053053 2020-02-01 21:58:52.399: INFO @log_variables: train gender_confidence/accuracy mean: 0.853831 2020-02-01 21:58:52.399: INFO @log_variables: train age_confidence/loss mean: 0.070835 2020-02-01 21:58:52.399: INFO @log_variables: train age_confidence/accuracy mean: 0.607092 2020-02-01 21:58:52.399: INFO @log_variables: valid loss nanmean: 0.842422 2020-02-01 21:58:52.399: INFO @log_variables: valid age_loss mean: 5.799661 2020-02-01 21:58:52.400: INFO @log_variables: valid gender_loss mean: 0.209263 2020-02-01 21:58:52.400: INFO @log_variables: valid age_mae mean: 6.278693 2020-02-01 21:58:52.400: INFO @log_variables: valid gender_accuracy mean: 0.925052 2020-02-01 21:58:52.400: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054330 2020-02-01 21:58:52.400: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875900 2020-02-01 21:58:52.400: INFO @log_variables: valid age_confidence/loss mean: 0.070611 2020-02-01 21:58:52.400: INFO @log_variables: valid age_confidence/accuracy mean: 0.560689 2020-02-01 21:58:52.400: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 21:58:52.407: INFO @metrics_hook: train age_mae: 5.539 +-0.032 (110372) 2020-02-01 21:58:52.414: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110372) 2020-02-01 21:58:55.164: INFO @metrics_hook: valid age_mae: 6.279 +-0.091 (17639) 2020-02-01 21:58:55.165: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 21:58:56.616: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:58:56.617: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 21:58:56.617: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 21:58:56.617: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 21:58:56.746: INFO @evaluate_confidence: Previous accuracy would be: 95.37 2020-02-01 21:58:56.746: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 21:58:56.806: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.03, 98.08, 98.16, 98.23, 98.3, 98.36, 98.41, 98.45, 98.5, 98.55, 98.59, 98.63, 98.67, 98.73, 98.77, 98.81, 98.84, 98.88, 98.92, 98.96, 99.0, 99.02, 99.06, 99.08, 99.11, 99.13, 99.16, 99.2, 99.23, 99.27, 99.29, 99.32, 99.35, 99.37, 99.39, 99.41, 99.45, 99.47, 99.5, 99.52, 99.55, 99.57, 99.59, 99.61, 99.62, 99.63, 99.66] 2020-02-01 21:58:56.806: INFO @evaluate_confidence: Dropped ratios are: [10.79, 11.2, 11.66, 12.1, 12.52, 12.93, 13.4, 13.8, 14.26, 14.7, 15.1, 15.55, 15.95, 16.41, 16.85, 17.27, 17.69, 18.15, 18.6, 19.04, 19.48, 19.9, 20.38, 20.86, 21.31, 21.77, 22.23, 22.71, 23.22, 23.68, 24.15, 24.66, 25.18, 25.67, 26.2, 26.7, 27.26, 27.84, 28.47, 29.09, 29.7, 30.35, 31.0, 31.72, 32.37, 33.1, 33.85] 2020-02-01 21:58:56.856: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:58:56.856: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 21:58:56.856: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 21:58:56.857: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 21:58:56.991: INFO @evaluate_confidence: Previous accuracy would be: 58.39 2020-02-01 21:58:56.991: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 21:58:57.006: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.58, 67.32, 67.95, 68.66, 69.44, 70.19, 71.15] 2020-02-01 21:58:57.006: INFO @evaluate_confidence: Dropped ratios are: [42.39, 45.63, 48.6, 51.74, 54.8, 57.65, 60.34] 2020-02-01 21:58:57.014: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 21:58:57.014: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 21:58:57.014: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.24 2020-02-01 21:58:57.014: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.24 2020-02-01 21:58:57.116: INFO @evaluate_confidence: Previous accuracy would be: 92.51 2020-02-01 21:58:57.117: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 21:58:57.125: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.49, 96.57, 96.68, 96.8, 96.9, 96.94, 96.96, 97.02, 97.09, 97.2, 97.28, 97.32, 97.35, 97.41, 97.46, 97.53, 97.56, 97.64, 97.72, 97.74, 97.75, 97.82, 97.89, 97.92, 97.99, 98.06, 98.14, 98.19, 98.25, 98.36, 98.42, 98.5, 98.56, 98.6, 98.67, 98.73, 98.77, 98.83, 98.88, 98.91] 2020-02-01 21:58:57.125: INFO @evaluate_confidence: Dropped ratios are: [13.25, 13.63, 14.0, 14.43, 14.89, 15.26, 15.6, 15.91, 16.37, 16.78, 17.15, 17.59, 17.91, 18.43, 18.78, 19.16, 19.47, 19.87, 20.3, 20.74, 21.18, 21.7, 22.14, 22.55, 23.03, 23.48, 24.15, 24.73, 25.35, 25.95, 26.63, 27.14, 27.66, 28.24, 28.86, 29.55, 30.19, 31.01, 31.86, 32.8] 2020-02-01 21:58:57.132: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 21:58:57.133: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 21:58:57.133: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 21:58:57.133: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 21:58:57.258: INFO @evaluate_confidence: Previous accuracy would be: 53.39 2020-02-01 21:58:57.258: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 21:58:57.259: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.23, 59.08, 59.27, 59.75] 2020-02-01 21:58:57.259: INFO @evaluate_confidence: Dropped ratios are: [43.32, 47.92, 52.72, 57.71] 2020-02-01 21:58:57.313: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:58:58.014: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:58:58.099: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:58:58.568: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:58:58.645: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 21:58:59.336: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:58:59.419: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:58:59.421: INFO @evaluate_gender-age_model: groups 0 3.390289 1 3.920727 2 5.229605 3 5.577601 4 6.329624 5 6.154586 6 6.172758 7 6.970513 Name: errors, dtype: float64 2020-02-01 21:58:59.422: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 21:58:59.882: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 21:58:59.943: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 21:58:59.944: INFO @evaluate_gender-age_model: groups 0 6.478823 1 5.703593 2 5.884532 3 5.773727 4 7.406483 5 5.436387 6 7.159759 7 11.078492 Name: errors, dtype: float64 2020-02-01 21:59:00.129: INFO @decay_lr : LR updated to `5.2119536e-05` 2020-02-01 21:59:00.131: INFO @log_profile : T train: 122.829621 2020-02-01 21:59:00.131: INFO @log_profile : T valid: 5.437415 2020-02-01 21:59:00.131: INFO @log_profile : T read data: 2.850899 2020-02-01 21:59:00.131: INFO @log_profile : T hooks: 10.223185 2020-02-01 21:59:00.131: INFO @main_loop : Epoch 130 done 2020-02-01 21:59:00.131: INFO @main_loop : Training epoch 131 2020-02-01 22:01:11.827: INFO @log_variables: train loss nanmean: 0.692618 2020-02-01 22:01:11.827: INFO @log_variables: train age_loss mean: 5.049655 2020-02-01 22:01:11.827: INFO @log_variables: train gender_loss mean: 0.120805 2020-02-01 22:01:11.827: INFO @log_variables: train age_mae mean: 5.524720 2020-02-01 22:01:11.827: INFO @log_variables: train gender_accuracy mean: 0.952678 2020-02-01 22:01:11.827: INFO @log_variables: train gender_confidence/loss nanmean: 0.052967 2020-02-01 22:01:11.827: INFO @log_variables: train gender_confidence/accuracy mean: 0.851711 2020-02-01 22:01:11.827: INFO @log_variables: train age_confidence/loss mean: 0.070768 2020-02-01 22:01:11.827: INFO @log_variables: train age_confidence/accuracy mean: 0.607310 2020-02-01 22:01:11.827: INFO @log_variables: valid loss nanmean: 0.827530 2020-02-01 22:01:11.827: INFO @log_variables: valid age_loss mean: 5.682455 2020-02-01 22:01:11.827: INFO @log_variables: valid gender_loss mean: 0.205481 2020-02-01 22:01:11.827: INFO @log_variables: valid age_mae mean: 6.162083 2020-02-01 22:01:11.827: INFO @log_variables: valid gender_accuracy mean: 0.923238 2020-02-01 22:01:11.827: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053387 2020-02-01 22:01:11.827: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868190 2020-02-01 22:01:11.828: INFO @log_variables: valid age_confidence/loss mean: 0.070755 2020-02-01 22:01:11.828: INFO @log_variables: valid age_confidence/accuracy mean: 0.554340 2020-02-01 22:01:11.828: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:01:11.835: INFO @metrics_hook: train age_mae: 5.525 +-0.032 (110372) 2020-02-01 22:01:11.842: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110372) 2020-02-01 22:01:14.557: INFO @metrics_hook: valid age_mae: 6.162 +-0.087 (17639) 2020-02-01 22:01:14.558: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:01:16.222: INFO @decay_lr : LR updated to `5.1858937e-05` 2020-02-01 22:01:16.223: INFO @log_profile : T train: 122.630891 2020-02-01 22:01:16.223: INFO @log_profile : T valid: 5.489396 2020-02-01 22:01:16.223: INFO @log_profile : T read data: 2.870350 2020-02-01 22:01:16.223: INFO @log_profile : T hooks: 5.024850 2020-02-01 22:01:16.223: INFO @main_loop : Epoch 131 done 2020-02-01 22:01:16.223: INFO @main_loop : Training epoch 132 2020-02-01 22:03:26.041: INFO @log_variables: train loss nanmean: 0.693152 2020-02-01 22:03:26.041: INFO @log_variables: train age_loss mean: 5.088645 2020-02-01 22:03:26.041: INFO @log_variables: train gender_loss mean: 0.118700 2020-02-01 22:03:26.041: INFO @log_variables: train age_mae mean: 5.564197 2020-02-01 22:03:26.041: INFO @log_variables: train gender_accuracy mean: 0.952863 2020-02-01 22:03:26.041: INFO @log_variables: train gender_confidence/loss nanmean: 0.052019 2020-02-01 22:03:26.041: INFO @log_variables: train gender_confidence/accuracy mean: 0.855152 2020-02-01 22:03:26.041: INFO @log_variables: train age_confidence/loss mean: 0.070619 2020-02-01 22:03:26.041: INFO @log_variables: train age_confidence/accuracy mean: 0.608190 2020-02-01 22:03:26.041: INFO @log_variables: valid loss nanmean: 0.833390 2020-02-01 22:03:26.041: INFO @log_variables: valid age_loss mean: 5.754062 2020-02-01 22:03:26.041: INFO @log_variables: valid gender_loss mean: 0.204185 2020-02-01 22:03:26.042: INFO @log_variables: valid age_mae mean: 6.233239 2020-02-01 22:03:26.042: INFO @log_variables: valid gender_accuracy mean: 0.923068 2020-02-01 22:03:26.042: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053981 2020-02-01 22:03:26.042: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869664 2020-02-01 22:03:26.042: INFO @log_variables: valid age_confidence/loss mean: 0.070690 2020-02-01 22:03:26.042: INFO @log_variables: valid age_confidence/accuracy mean: 0.557401 2020-02-01 22:03:26.042: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:03:26.049: INFO @metrics_hook: train age_mae: 5.564 +-0.032 (110592) 2020-02-01 22:03:26.056: INFO @metrics_hook: train gender_accuracy: 0.953 +-0.001 (110592) 2020-02-01 22:03:28.846: INFO @metrics_hook: valid age_mae: 6.233 +-0.088 (17639) 2020-02-01 22:03:28.847: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:03:30.291: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:03:30.291: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:03:30.291: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.20 2020-02-01 22:03:30.291: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:03:30.422: INFO @evaluate_confidence: Previous accuracy would be: 95.29 2020-02-01 22:03:30.423: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:03:30.483: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.03, 98.1, 98.17, 98.25, 98.32, 98.39, 98.45, 98.5, 98.54, 98.61, 98.65, 98.7, 98.75, 98.8, 98.84, 98.88, 98.92, 98.95, 98.98, 99.03, 99.05, 99.09, 99.13, 99.16, 99.2, 99.24, 99.26, 99.29, 99.31, 99.35, 99.37, 99.39, 99.41, 99.43, 99.45, 99.47, 99.48, 99.5, 99.52, 99.54, 99.56, 99.57, 99.6, 99.62, 99.63, 99.66, 99.67] 2020-02-01 22:03:30.483: INFO @evaluate_confidence: Dropped ratios are: [10.9, 11.35, 11.79, 12.23, 12.68, 13.11, 13.54, 13.95, 14.35, 14.79, 15.22, 15.66, 16.1, 16.53, 16.94, 17.34, 17.75, 18.18, 18.61, 19.04, 19.46, 19.89, 20.31, 20.77, 21.22, 21.72, 22.18, 22.65, 23.14, 23.63, 24.12, 24.64, 25.18, 25.7, 26.22, 26.76, 27.3, 27.9, 28.46, 29.08, 29.68, 30.32, 30.99, 31.64, 32.35, 33.06, 33.8] 2020-02-01 22:03:30.532: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:03:30.532: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:03:30.532: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:03:30.532: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 22:03:30.665: INFO @evaluate_confidence: Previous accuracy would be: 58.16 2020-02-01 22:03:30.665: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:03:30.680: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.51, 67.19, 67.9, 68.64, 69.37, 70.19, 71.11] 2020-02-01 22:03:30.680: INFO @evaluate_confidence: Dropped ratios are: [42.55, 45.78, 48.85, 51.87, 54.83, 57.73, 60.39] 2020-02-01 22:03:30.687: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:03:30.688: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 22:03:30.688: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.24 2020-02-01 22:03:30.688: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:03:30.790: INFO @evaluate_confidence: Previous accuracy would be: 92.31 2020-02-01 22:03:30.790: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:03:30.799: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.54, 96.61, 96.69, 96.76, 96.83, 96.86, 96.97, 97.04, 97.15, 97.19, 97.23, 97.31, 97.39, 97.44, 97.48, 97.55, 97.6, 97.66, 97.74, 97.76, 97.81, 97.86, 97.9, 97.94, 97.98, 98.03, 98.1, 98.15, 98.24, 98.32, 98.34, 98.39, 98.42, 98.49, 98.54, 98.56, 98.68, 98.71, 98.74, 98.75, 98.82] 2020-02-01 22:03:30.799: INFO @evaluate_confidence: Dropped ratios are: [13.57, 13.95, 14.29, 14.67, 14.97, 15.42, 15.85, 16.21, 16.59, 16.98, 17.36, 17.8, 18.22, 18.66, 18.99, 19.4, 19.98, 20.38, 20.8, 21.23, 21.65, 22.09, 22.61, 23.06, 23.48, 24.05, 24.55, 25.0, 25.51, 26.04, 26.69, 27.25, 27.75, 28.44, 29.13, 29.7, 30.44, 31.18, 31.88, 32.73, 33.52] 2020-02-01 22:03:30.807: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:03:30.807: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 22:03:30.807: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 22:03:30.807: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:03:30.931: INFO @evaluate_confidence: Previous accuracy would be: 53.16 2020-02-01 22:03:30.931: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 22:03:30.932: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.52, 58.88, 59.34] 2020-02-01 22:03:30.932: INFO @evaluate_confidence: Dropped ratios are: [47.8, 52.24, 56.8] 2020-02-01 22:03:30.986: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:03:31.682: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:03:31.767: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:03:32.252: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:03:32.510: INFO @decay_lr : LR updated to `5.1599643e-05` 2020-02-01 22:03:32.512: INFO @log_profile : T train: 121.812990 2020-02-01 22:03:32.512: INFO @log_profile : T valid: 5.424228 2020-02-01 22:03:32.512: INFO @log_profile : T read data: 1.882999 2020-02-01 22:03:32.512: INFO @log_profile : T hooks: 7.089447 2020-02-01 22:03:32.512: INFO @main_loop : Epoch 132 done 2020-02-01 22:03:32.512: INFO @main_loop : Training epoch 133 2020-02-01 22:05:43.245: INFO @log_variables: train loss nanmean: 0.691412 2020-02-01 22:05:43.246: INFO @log_variables: train age_loss mean: 5.054744 2020-02-01 22:05:43.246: INFO @log_variables: train gender_loss mean: 0.118745 2020-02-01 22:05:43.246: INFO @log_variables: train age_mae mean: 5.530169 2020-02-01 22:05:43.246: INFO @log_variables: train gender_accuracy mean: 0.954200 2020-02-01 22:05:43.246: INFO @log_variables: train gender_confidence/loss nanmean: 0.053111 2020-02-01 22:05:43.246: INFO @log_variables: train gender_confidence/accuracy mean: 0.852463 2020-02-01 22:05:43.246: INFO @log_variables: train age_confidence/loss mean: 0.070829 2020-02-01 22:05:43.246: INFO @log_variables: train age_confidence/accuracy mean: 0.607165 2020-02-01 22:05:43.246: INFO @log_variables: valid loss nanmean: 0.845436 2020-02-01 22:05:43.246: INFO @log_variables: valid age_loss mean: 5.862593 2020-02-01 22:05:43.246: INFO @log_variables: valid gender_loss mean: 0.207865 2020-02-01 22:05:43.246: INFO @log_variables: valid age_mae mean: 6.343446 2020-02-01 22:05:43.246: INFO @log_variables: valid gender_accuracy mean: 0.921991 2020-02-01 22:05:43.246: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054490 2020-02-01 22:05:43.246: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866319 2020-02-01 22:05:43.246: INFO @log_variables: valid age_confidence/loss mean: 0.069015 2020-02-01 22:05:43.246: INFO @log_variables: valid age_confidence/accuracy mean: 0.561483 2020-02-01 22:05:43.246: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:05:43.254: INFO @metrics_hook: train age_mae: 5.530 +-0.032 (110372) 2020-02-01 22:05:43.261: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110372) 2020-02-01 22:05:46.062: INFO @metrics_hook: valid age_mae: 6.343 +-0.088 (17639) 2020-02-01 22:05:46.064: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 22:05:47.758: INFO @decay_lr : LR updated to `5.1341645e-05` 2020-02-01 22:05:47.760: INFO @log_profile : T train: 121.733938 2020-02-01 22:05:47.760: INFO @log_profile : T valid: 5.423132 2020-02-01 22:05:47.760: INFO @log_profile : T read data: 2.879611 2020-02-01 22:05:47.760: INFO @log_profile : T hooks: 5.133936 2020-02-01 22:05:47.760: INFO @main_loop : Epoch 133 done 2020-02-01 22:05:47.760: INFO @main_loop : Training epoch 134 2020-02-01 22:07:58.337: INFO @log_variables: train loss nanmean: 0.688418 2020-02-01 22:07:58.337: INFO @log_variables: train age_loss mean: 5.034648 2020-02-01 22:07:58.337: INFO @log_variables: train gender_loss mean: 0.118022 2020-02-01 22:07:58.337: INFO @log_variables: train age_mae mean: 5.509545 2020-02-01 22:07:58.337: INFO @log_variables: train gender_accuracy mean: 0.953584 2020-02-01 22:07:58.337: INFO @log_variables: train gender_confidence/loss nanmean: 0.052593 2020-02-01 22:07:58.337: INFO @log_variables: train gender_confidence/accuracy mean: 0.855326 2020-02-01 22:07:58.337: INFO @log_variables: train age_confidence/loss mean: 0.070838 2020-02-01 22:07:58.337: INFO @log_variables: train age_confidence/accuracy mean: 0.609276 2020-02-01 22:07:58.338: INFO @log_variables: valid loss nanmean: 0.829245 2020-02-01 22:07:58.338: INFO @log_variables: valid age_loss mean: 5.781368 2020-02-01 22:07:58.338: INFO @log_variables: valid gender_loss mean: 0.195323 2020-02-01 22:07:58.338: INFO @log_variables: valid age_mae mean: 6.262008 2020-02-01 22:07:58.338: INFO @log_variables: valid gender_accuracy mean: 0.926697 2020-02-01 22:07:58.338: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055750 2020-02-01 22:07:58.338: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873972 2020-02-01 22:07:58.338: INFO @log_variables: valid age_confidence/loss mean: 0.070349 2020-02-01 22:07:58.338: INFO @log_variables: valid age_confidence/accuracy mean: 0.553830 2020-02-01 22:07:58.338: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:07:58.345: INFO @metrics_hook: train age_mae: 5.510 +-0.032 (110372) 2020-02-01 22:07:58.352: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110372) 2020-02-01 22:08:01.121: INFO @metrics_hook: valid age_mae: 6.262 +-0.090 (17639) 2020-02-01 22:08:01.122: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 22:08:02.606: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:08:02.606: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:08:02.606: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 22:08:02.606: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:08:02.739: INFO @evaluate_confidence: Previous accuracy would be: 95.36 2020-02-01 22:08:02.740: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:08:02.801: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.02, 98.09, 98.16, 98.22, 98.26, 98.32, 98.38, 98.47, 98.52, 98.59, 98.63, 98.67, 98.71, 98.75, 98.8, 98.85, 98.89, 98.93, 98.98, 99.02, 99.04, 99.07, 99.11, 99.14, 99.17, 99.2, 99.23, 99.26, 99.29, 99.32, 99.35, 99.36, 99.38, 99.41, 99.43, 99.44, 99.46, 99.49, 99.51, 99.53, 99.54, 99.56, 99.58, 99.59, 99.61, 99.63, 99.65] 2020-02-01 22:08:02.801: INFO @evaluate_confidence: Dropped ratios are: [10.82, 11.25, 11.65, 12.09, 12.49, 12.93, 13.34, 13.77, 14.17, 14.65, 15.05, 15.49, 15.91, 16.33, 16.78, 17.21, 17.63, 18.11, 18.58, 19.02, 19.48, 19.92, 20.37, 20.82, 21.28, 21.72, 22.18, 22.68, 23.16, 23.65, 24.16, 24.67, 25.16, 25.71, 26.24, 26.79, 27.32, 27.89, 28.51, 29.11, 29.73, 30.36, 31.0, 31.63, 32.35, 33.12, 33.92] 2020-02-01 22:08:02.851: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:08:02.851: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:08:02.852: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:08:02.852: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:08:02.993: INFO @evaluate_confidence: Previous accuracy would be: 58.47 2020-02-01 22:08:02.993: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:08:03.008: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.64, 67.26, 67.85, 68.56, 69.41, 70.21, 71.09] 2020-02-01 22:08:03.008: INFO @evaluate_confidence: Dropped ratios are: [41.73, 44.93, 48.15, 51.28, 54.36, 57.21, 59.98] 2020-02-01 22:08:03.016: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:08:03.016: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 22:08:03.016: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.24 2020-02-01 22:08:03.016: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.24 2020-02-01 22:08:03.122: INFO @evaluate_confidence: Previous accuracy would be: 92.67 2020-02-01 22:08:03.122: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:08:03.131: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.43, 96.55, 96.63, 96.73, 96.81, 96.87, 96.93, 97.0, 97.09, 97.13, 97.22, 97.29, 97.39, 97.45, 97.5, 97.6, 97.68, 97.77, 97.85, 97.9, 97.98, 98.03, 98.08, 98.13, 98.19, 98.29, 98.32, 98.43, 98.47, 98.51, 98.55, 98.58, 98.62, 98.66, 98.77, 98.81, 98.87, 98.91, 98.94, 98.98, 99.0] 2020-02-01 22:08:03.131: INFO @evaluate_confidence: Dropped ratios are: [12.62, 13.02, 13.49, 13.84, 14.25, 14.59, 14.97, 15.4, 15.78, 16.25, 16.71, 17.08, 17.59, 18.02, 18.37, 18.86, 19.32, 19.78, 20.24, 20.6, 20.99, 21.38, 21.85, 22.34, 22.88, 23.36, 23.74, 24.32, 24.72, 25.19, 25.81, 26.34, 26.94, 27.59, 28.24, 29.0, 29.71, 30.5, 31.27, 32.1, 32.93] 2020-02-01 22:08:03.139: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:08:03.139: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.52 +- 0.11 2020-02-01 22:08:03.139: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.48 +- 0.10 2020-02-01 22:08:03.139: INFO @evaluate_confidence: Average confidence of all samples 0.50 +- 0.10 2020-02-01 22:08:03.270: INFO @evaluate_confidence: Previous accuracy would be: 52.96 2020-02-01 22:08:03.270: INFO @evaluate_confidence: Possible optimal thresholds are: [0.48, 0.49, 0.5, 0.51] 2020-02-01 22:08:03.272: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.45, 58.82, 59.25, 59.85] 2020-02-01 22:08:03.272: INFO @evaluate_confidence: Dropped ratios are: [45.34, 49.89, 54.9, 59.41] 2020-02-01 22:08:03.327: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:08:04.012: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:08:04.098: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:08:04.560: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:08:04.823: INFO @decay_lr : LR updated to `5.108494e-05` 2020-02-01 22:08:04.825: INFO @log_profile : T train: 121.536938 2020-02-01 22:08:04.825: INFO @log_profile : T valid: 5.490839 2020-02-01 22:08:04.825: INFO @log_profile : T read data: 2.842578 2020-02-01 22:08:04.825: INFO @log_profile : T hooks: 7.117538 2020-02-01 22:08:04.825: INFO @main_loop : Epoch 134 done 2020-02-01 22:08:04.825: INFO @main_loop : Training epoch 135 2020-02-01 22:10:14.812: INFO @log_variables: train loss nanmean: 0.690342 2020-02-01 22:10:14.812: INFO @log_variables: train age_loss mean: 5.063212 2020-02-01 22:10:14.812: INFO @log_variables: train gender_loss mean: 0.117252 2020-02-01 22:10:14.812: INFO @log_variables: train age_mae mean: 5.538731 2020-02-01 22:10:14.812: INFO @log_variables: train gender_accuracy mean: 0.954210 2020-02-01 22:10:14.812: INFO @log_variables: train gender_confidence/loss nanmean: 0.052547 2020-02-01 22:10:14.812: INFO @log_variables: train gender_confidence/accuracy mean: 0.854411 2020-02-01 22:10:14.812: INFO @log_variables: train age_confidence/loss mean: 0.070911 2020-02-01 22:10:14.812: INFO @log_variables: train age_confidence/accuracy mean: 0.606174 2020-02-01 22:10:14.812: INFO @log_variables: valid loss nanmean: 0.847319 2020-02-01 22:10:14.812: INFO @log_variables: valid age_loss mean: 5.815017 2020-02-01 22:10:14.812: INFO @log_variables: valid gender_loss mean: 0.211560 2020-02-01 22:10:14.812: INFO @log_variables: valid age_mae mean: 6.294442 2020-02-01 22:10:14.812: INFO @log_variables: valid gender_accuracy mean: 0.924089 2020-02-01 22:10:14.813: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055770 2020-02-01 22:10:14.813: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873462 2020-02-01 22:10:14.813: INFO @log_variables: valid age_confidence/loss mean: 0.070583 2020-02-01 22:10:14.813: INFO @log_variables: valid age_confidence/accuracy mean: 0.572368 2020-02-01 22:10:14.813: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:10:14.820: INFO @metrics_hook: train age_mae: 5.539 +-0.032 (110592) 2020-02-01 22:10:14.827: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110592) 2020-02-01 22:10:17.551: INFO @metrics_hook: valid age_mae: 6.294 +-0.091 (17639) 2020-02-01 22:10:17.552: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 22:10:19.196: INFO @decay_lr : LR updated to `5.0829512e-05` 2020-02-01 22:10:19.198: INFO @log_profile : T train: 121.936485 2020-02-01 22:10:19.198: INFO @log_profile : T valid: 5.465383 2020-02-01 22:10:19.198: INFO @log_profile : T read data: 1.876063 2020-02-01 22:10:19.198: INFO @log_profile : T hooks: 5.018048 2020-02-01 22:10:19.198: INFO @main_loop : Epoch 135 done 2020-02-01 22:10:19.198: INFO @main_loop : Training epoch 136 2020-02-01 22:12:29.799: INFO @log_variables: train loss nanmean: 0.686949 2020-02-01 22:12:29.799: INFO @log_variables: train age_loss mean: 5.020851 2020-02-01 22:12:29.799: INFO @log_variables: train gender_loss mean: 0.116991 2020-02-01 22:12:29.799: INFO @log_variables: train age_mae mean: 5.496141 2020-02-01 22:12:29.799: INFO @log_variables: train gender_accuracy mean: 0.954617 2020-02-01 22:12:29.799: INFO @log_variables: train gender_confidence/loss nanmean: 0.053382 2020-02-01 22:12:29.799: INFO @log_variables: train gender_confidence/accuracy mean: 0.852716 2020-02-01 22:12:29.799: INFO @log_variables: train age_confidence/loss mean: 0.070770 2020-02-01 22:12:29.799: INFO @log_variables: train age_confidence/accuracy mean: 0.610436 2020-02-01 22:12:29.799: INFO @log_variables: valid loss nanmean: 0.841834 2020-02-01 22:12:29.799: INFO @log_variables: valid age_loss mean: 5.861794 2020-02-01 22:12:29.799: INFO @log_variables: valid gender_loss mean: 0.202316 2020-02-01 22:12:29.799: INFO @log_variables: valid age_mae mean: 6.341621 2020-02-01 22:12:29.800: INFO @log_variables: valid gender_accuracy mean: 0.922558 2020-02-01 22:12:29.800: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054933 2020-02-01 22:12:29.800: INFO @log_variables: valid gender_confidence/accuracy mean: 0.861387 2020-02-01 22:12:29.800: INFO @log_variables: valid age_confidence/loss mean: 0.070087 2020-02-01 22:12:29.800: INFO @log_variables: valid age_confidence/accuracy mean: 0.565452 2020-02-01 22:12:29.800: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:12:29.807: INFO @metrics_hook: train age_mae: 5.496 +-0.032 (110372) 2020-02-01 22:12:29.815: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110372) 2020-02-01 22:12:32.557: INFO @metrics_hook: valid age_mae: 6.342 +-0.091 (17639) 2020-02-01 22:12:32.559: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:12:34.008: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:12:34.008: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:12:34.008: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 22:12:34.009: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:12:34.137: INFO @evaluate_confidence: Previous accuracy would be: 95.46 2020-02-01 22:12:34.138: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:12:34.198: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.07, 98.13, 98.19, 98.25, 98.32, 98.37, 98.43, 98.49, 98.54, 98.59, 98.64, 98.67, 98.72, 98.76, 98.8, 98.84, 98.88, 98.92, 98.96, 98.99, 99.02, 99.05, 99.09, 99.11, 99.15, 99.18, 99.2, 99.23, 99.27, 99.29, 99.31, 99.33, 99.36, 99.38, 99.41, 99.44, 99.46, 99.48, 99.5, 99.52, 99.53, 99.55, 99.56, 99.57, 99.59, 99.61, 99.62] 2020-02-01 22:12:34.198: INFO @evaluate_confidence: Dropped ratios are: [10.91, 11.36, 11.81, 12.21, 12.64, 13.07, 13.52, 13.93, 14.33, 14.78, 15.2, 15.6, 16.03, 16.46, 16.9, 17.35, 17.75, 18.17, 18.65, 19.06, 19.5, 19.94, 20.37, 20.81, 21.26, 21.73, 22.24, 22.73, 23.25, 23.71, 24.23, 24.71, 25.19, 25.71, 26.23, 26.83, 27.38, 27.99, 28.57, 29.22, 29.83, 30.46, 31.12, 31.79, 32.52, 33.23, 33.98] 2020-02-01 22:12:34.247: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:12:34.248: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:12:34.248: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:12:34.248: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 22:12:34.384: INFO @evaluate_confidence: Previous accuracy would be: 58.64 2020-02-01 22:12:34.384: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:12:34.399: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.99, 67.57, 68.37, 69.06, 69.89, 70.63, 71.62] 2020-02-01 22:12:34.399: INFO @evaluate_confidence: Dropped ratios are: [42.08, 45.16, 48.28, 51.31, 54.32, 57.16, 59.95] 2020-02-01 22:12:34.406: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:12:34.407: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.23 2020-02-01 22:12:34.407: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.23 2020-02-01 22:12:34.407: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 22:12:34.509: INFO @evaluate_confidence: Previous accuracy would be: 92.26 2020-02-01 22:12:34.509: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 22:12:34.517: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.4, 96.5, 96.62, 96.72, 96.81, 96.91, 96.97, 97.08, 97.17, 97.25, 97.3, 97.42, 97.46, 97.53, 97.6, 97.67, 97.73, 97.82, 97.9, 97.94, 98.02, 98.13, 98.23, 98.28, 98.32, 98.38, 98.44, 98.5, 98.56, 98.61, 98.65, 98.7, 98.73, 98.77, 98.81, 98.83, 98.86, 98.88, 98.92, 98.95] 2020-02-01 22:12:34.517: INFO @evaluate_confidence: Dropped ratios are: [13.8, 14.25, 14.73, 15.18, 15.66, 16.17, 16.55, 17.01, 17.36, 17.76, 18.23, 18.75, 19.22, 19.64, 20.09, 20.53, 21.03, 21.52, 22.07, 22.57, 23.11, 23.65, 24.24, 24.66, 25.18, 25.73, 26.25, 26.84, 27.41, 28.02, 28.56, 29.17, 29.84, 30.39, 31.12, 31.76, 32.41, 33.11, 33.9, 34.58] 2020-02-01 22:12:34.525: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:12:34.525: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 22:12:34.525: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 22:12:34.525: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:12:34.649: INFO @evaluate_confidence: Previous accuracy would be: 52.50 2020-02-01 22:12:34.649: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 22:12:34.651: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.77, 59.21, 59.82, 60.55] 2020-02-01 22:12:34.651: INFO @evaluate_confidence: Dropped ratios are: [46.35, 50.89, 55.27, 59.69] 2020-02-01 22:12:34.702: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:12:35.385: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:12:35.467: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:12:35.934: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:12:36.192: INFO @decay_lr : LR updated to `5.0575363e-05` 2020-02-01 22:12:36.193: INFO @log_profile : T train: 121.543079 2020-02-01 22:12:36.193: INFO @log_profile : T valid: 5.467715 2020-02-01 22:12:36.193: INFO @log_profile : T read data: 2.930405 2020-02-01 22:12:36.193: INFO @log_profile : T hooks: 6.975760 2020-02-01 22:12:36.193: INFO @main_loop : Epoch 136 done 2020-02-01 22:12:36.193: INFO @main_loop : Training epoch 137 2020-02-01 22:14:47.144: INFO @log_variables: train loss nanmean: 0.683794 2020-02-01 22:14:47.144: INFO @log_variables: train age_loss mean: 5.019215 2020-02-01 22:14:47.144: INFO @log_variables: train gender_loss mean: 0.115293 2020-02-01 22:14:47.145: INFO @log_variables: train age_mae mean: 5.494936 2020-02-01 22:14:47.145: INFO @log_variables: train gender_accuracy mean: 0.954463 2020-02-01 22:14:47.145: INFO @log_variables: train gender_confidence/loss nanmean: 0.051747 2020-02-01 22:14:47.145: INFO @log_variables: train gender_confidence/accuracy mean: 0.853767 2020-02-01 22:14:47.145: INFO @log_variables: train age_confidence/loss mean: 0.070942 2020-02-01 22:14:47.145: INFO @log_variables: train age_confidence/accuracy mean: 0.610716 2020-02-01 22:14:47.145: INFO @log_variables: valid loss nanmean: 0.842383 2020-02-01 22:14:47.145: INFO @log_variables: valid age_loss mean: 5.832016 2020-02-01 22:14:47.145: INFO @log_variables: valid gender_loss mean: 0.205736 2020-02-01 22:14:47.145: INFO @log_variables: valid age_mae mean: 6.311589 2020-02-01 22:14:47.145: INFO @log_variables: valid gender_accuracy mean: 0.924712 2020-02-01 22:14:47.145: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054683 2020-02-01 22:14:47.145: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875390 2020-02-01 22:14:47.145: INFO @log_variables: valid age_confidence/loss mean: 0.070484 2020-02-01 22:14:47.145: INFO @log_variables: valid age_confidence/accuracy mean: 0.552979 2020-02-01 22:14:47.145: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:14:47.153: INFO @metrics_hook: train age_mae: 5.495 +-0.032 (110372) 2020-02-01 22:14:47.160: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110372) 2020-02-01 22:14:49.858: INFO @metrics_hook: valid age_mae: 6.312 +-0.091 (17639) 2020-02-01 22:14:49.860: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 22:14:51.509: INFO @decay_lr : LR updated to `5.0322487e-05` 2020-02-01 22:14:51.510: INFO @log_profile : T train: 121.849311 2020-02-01 22:14:51.510: INFO @log_profile : T valid: 5.514101 2020-02-01 22:14:51.510: INFO @log_profile : T read data: 2.869172 2020-02-01 22:14:51.510: INFO @log_profile : T hooks: 5.007259 2020-02-01 22:14:51.510: INFO @main_loop : Epoch 137 done 2020-02-01 22:14:51.511: INFO @main_loop : Training epoch 138 2020-02-01 22:17:01.431: INFO @log_variables: train loss nanmean: 0.683898 2020-02-01 22:17:01.431: INFO @log_variables: train age_loss mean: 5.015215 2020-02-01 22:17:01.431: INFO @log_variables: train gender_loss mean: 0.115825 2020-02-01 22:17:01.432: INFO @log_variables: train age_mae mean: 5.490393 2020-02-01 22:17:01.432: INFO @log_variables: train gender_accuracy mean: 0.954834 2020-02-01 22:17:01.432: INFO @log_variables: train gender_confidence/loss nanmean: 0.051819 2020-02-01 22:17:01.432: INFO @log_variables: train gender_confidence/accuracy mean: 0.855812 2020-02-01 22:17:01.432: INFO @log_variables: train age_confidence/loss mean: 0.070855 2020-02-01 22:17:01.432: INFO @log_variables: train age_confidence/accuracy mean: 0.609257 2020-02-01 22:17:01.432: INFO @log_variables: valid loss nanmean: 0.828781 2020-02-01 22:17:01.432: INFO @log_variables: valid age_loss mean: 5.802912 2020-02-01 22:17:01.432: INFO @log_variables: valid gender_loss mean: 0.194761 2020-02-01 22:17:01.432: INFO @log_variables: valid age_mae mean: 6.282363 2020-02-01 22:17:01.432: INFO @log_variables: valid gender_accuracy mean: 0.926526 2020-02-01 22:17:01.432: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054548 2020-02-01 22:17:01.432: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877487 2020-02-01 22:17:01.432: INFO @log_variables: valid age_confidence/loss mean: 0.069641 2020-02-01 22:17:01.432: INFO @log_variables: valid age_confidence/accuracy mean: 0.556834 2020-02-01 22:17:01.432: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:17:01.439: INFO @metrics_hook: train age_mae: 5.490 +-0.032 (110592) 2020-02-01 22:17:01.446: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110592) 2020-02-01 22:17:04.194: INFO @metrics_hook: valid age_mae: 6.282 +-0.089 (17639) 2020-02-01 22:17:04.195: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 22:17:05.659: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:17:05.659: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:17:05.659: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:17:05.660: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:17:05.789: INFO @evaluate_confidence: Previous accuracy would be: 95.48 2020-02-01 22:17:05.789: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:17:05.851: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.04, 98.12, 98.19, 98.26, 98.34, 98.4, 98.46, 98.51, 98.56, 98.61, 98.66, 98.7, 98.74, 98.78, 98.83, 98.87, 98.9, 98.93, 98.97, 99.01, 99.05, 99.09, 99.12, 99.15, 99.19, 99.21, 99.24, 99.27, 99.29, 99.32, 99.33, 99.37, 99.39, 99.41, 99.43, 99.46, 99.47, 99.5, 99.51, 99.54, 99.57, 99.59, 99.61, 99.61, 99.63, 99.64, 99.66, 99.68] 2020-02-01 22:17:05.851: INFO @evaluate_confidence: Dropped ratios are: [10.34, 10.78, 11.21, 11.64, 12.09, 12.5, 12.93, 13.36, 13.78, 14.22, 14.64, 15.06, 15.47, 15.86, 16.28, 16.69, 17.1, 17.53, 17.95, 18.39, 18.83, 19.27, 19.7, 20.12, 20.56, 20.99, 21.42, 21.88, 22.3, 22.78, 23.25, 23.74, 24.27, 24.79, 25.31, 25.84, 26.36, 26.91, 27.43, 27.99, 28.6, 29.21, 29.78, 30.35, 31.02, 31.73, 32.47, 33.25] 2020-02-01 22:17:05.899: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:17:05.900: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:17:05.900: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:17:05.900: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.14 2020-02-01 22:17:06.033: INFO @evaluate_confidence: Previous accuracy would be: 58.71 2020-02-01 22:17:06.034: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:17:06.049: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.96, 67.58, 68.29, 69.02, 69.73, 70.56, 71.3] 2020-02-01 22:17:06.049: INFO @evaluate_confidence: Dropped ratios are: [42.09, 45.15, 48.28, 51.34, 54.28, 57.12, 59.77] 2020-02-01 22:17:06.056: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:17:06.056: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 22:17:06.056: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.24 2020-02-01 22:17:06.057: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.24 2020-02-01 22:17:06.161: INFO @evaluate_confidence: Previous accuracy would be: 92.65 2020-02-01 22:17:06.161: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 22:17:06.170: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.57, 96.64, 96.75, 96.86, 96.98, 97.06, 97.16, 97.24, 97.31, 97.37, 97.43, 97.5, 97.56, 97.62, 97.68, 97.72, 97.78, 97.83, 97.88, 97.96, 98.0, 98.02, 98.06, 98.1, 98.16, 98.23, 98.29, 98.34, 98.37, 98.4, 98.43, 98.49, 98.57, 98.64, 98.67, 98.72, 98.75, 98.79, 98.81, 98.87, 98.98] 2020-02-01 22:17:06.170: INFO @evaluate_confidence: Dropped ratios are: [12.8, 13.18, 13.61, 14.0, 14.43, 14.85, 15.27, 15.67, 16.07, 16.49, 16.95, 17.33, 17.78, 18.17, 18.55, 18.97, 19.45, 19.88, 20.31, 20.83, 21.31, 21.69, 22.12, 22.57, 23.09, 23.61, 24.16, 24.75, 25.24, 25.76, 26.24, 26.74, 27.42, 28.12, 28.75, 29.4, 30.18, 30.88, 31.65, 32.56, 33.47] 2020-02-01 22:17:06.177: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:17:06.178: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 22:17:06.178: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 22:17:06.178: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:17:06.303: INFO @evaluate_confidence: Previous accuracy would be: 52.04 2020-02-01 22:17:06.303: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 22:17:06.304: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.45, 57.57, 57.67, 58.32] 2020-02-01 22:17:06.304: INFO @evaluate_confidence: Dropped ratios are: [44.67, 48.97, 53.63, 58.21] 2020-02-01 22:17:06.357: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:17:07.044: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:17:07.130: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:17:07.592: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:17:07.851: INFO @decay_lr : LR updated to `5.0070874e-05` 2020-02-01 22:17:07.852: INFO @log_profile : T train: 121.796930 2020-02-01 22:17:07.852: INFO @log_profile : T valid: 5.479290 2020-02-01 22:17:07.852: INFO @log_profile : T read data: 1.933387 2020-02-01 22:17:07.852: INFO @log_profile : T hooks: 7.053089 2020-02-01 22:17:07.852: INFO @main_loop : Epoch 138 done 2020-02-01 22:17:07.852: INFO @main_loop : Training epoch 139 2020-02-01 22:19:18.507: INFO @log_variables: train loss nanmean: 0.681572 2020-02-01 22:19:18.507: INFO @log_variables: train age_loss mean: 4.982542 2020-02-01 22:19:18.507: INFO @log_variables: train gender_loss mean: 0.115729 2020-02-01 22:19:18.507: INFO @log_variables: train age_mae mean: 5.458009 2020-02-01 22:19:18.507: INFO @log_variables: train gender_accuracy mean: 0.954753 2020-02-01 22:19:18.507: INFO @log_variables: train gender_confidence/loss nanmean: 0.052305 2020-02-01 22:19:18.507: INFO @log_variables: train gender_confidence/accuracy mean: 0.855344 2020-02-01 22:19:18.507: INFO @log_variables: train age_confidence/loss mean: 0.071100 2020-02-01 22:19:18.507: INFO @log_variables: train age_confidence/accuracy mean: 0.607953 2020-02-01 22:19:18.507: INFO @log_variables: valid loss nanmean: 0.844190 2020-02-01 22:19:18.507: INFO @log_variables: valid age_loss mean: 5.869687 2020-02-01 22:19:18.507: INFO @log_variables: valid gender_loss mean: 0.206399 2020-02-01 22:19:18.507: INFO @log_variables: valid age_mae mean: 6.348913 2020-02-01 22:19:18.507: INFO @log_variables: valid gender_accuracy mean: 0.922728 2020-02-01 22:19:18.507: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053368 2020-02-01 22:19:18.508: INFO @log_variables: valid gender_confidence/accuracy mean: 0.865639 2020-02-01 22:19:18.508: INFO @log_variables: valid age_confidence/loss mean: 0.069579 2020-02-01 22:19:18.508: INFO @log_variables: valid age_confidence/accuracy mean: 0.557855 2020-02-01 22:19:18.508: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:19:18.515: INFO @metrics_hook: train age_mae: 5.458 +-0.031 (110372) 2020-02-01 22:19:18.523: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110372) 2020-02-01 22:19:21.206: INFO @metrics_hook: valid age_mae: 6.349 +-0.091 (17639) 2020-02-01 22:19:21.207: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:19:22.835: INFO @decay_lr : LR updated to `4.982052e-05` 2020-02-01 22:19:22.836: INFO @log_profile : T train: 121.636717 2020-02-01 22:19:22.836: INFO @log_profile : T valid: 5.462471 2020-02-01 22:19:22.836: INFO @log_profile : T read data: 2.846862 2020-02-01 22:19:22.836: INFO @log_profile : T hooks: 4.961654 2020-02-01 22:19:22.836: INFO @main_loop : Epoch 139 done 2020-02-01 22:19:22.836: INFO @main_loop : Training epoch 140 2020-02-01 22:21:43.842: INFO @log_variables: train loss nanmean: 0.681106 2020-02-01 22:21:43.842: INFO @log_variables: train age_loss mean: 4.997262 2020-02-01 22:21:43.842: INFO @log_variables: train gender_loss mean: 0.114168 2020-02-01 22:21:43.842: INFO @log_variables: train age_mae mean: 5.472292 2020-02-01 22:21:43.842: INFO @log_variables: train gender_accuracy mean: 0.955360 2020-02-01 22:21:43.842: INFO @log_variables: train gender_confidence/loss nanmean: 0.051899 2020-02-01 22:21:43.842: INFO @log_variables: train gender_confidence/accuracy mean: 0.857382 2020-02-01 22:21:43.842: INFO @log_variables: train age_confidence/loss mean: 0.071121 2020-02-01 22:21:43.843: INFO @log_variables: train age_confidence/accuracy mean: 0.611577 2020-02-01 22:21:43.843: INFO @log_variables: valid loss nanmean: 0.833080 2020-02-01 22:21:43.843: INFO @log_variables: valid age_loss mean: 5.812181 2020-02-01 22:21:43.843: INFO @log_variables: valid gender_loss mean: 0.198918 2020-02-01 22:21:43.843: INFO @log_variables: valid age_mae mean: 6.291534 2020-02-01 22:21:43.843: INFO @log_variables: valid gender_accuracy mean: 0.926073 2020-02-01 22:21:43.843: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054068 2020-02-01 22:21:43.843: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871988 2020-02-01 22:21:43.843: INFO @log_variables: valid age_confidence/loss mean: 0.069797 2020-02-01 22:21:43.843: INFO @log_variables: valid age_confidence/accuracy mean: 0.557231 2020-02-01 22:21:43.843: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:21:43.850: INFO @metrics_hook: train age_mae: 5.472 +-0.032 (110372) 2020-02-01 22:21:43.857: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110372) 2020-02-01 22:21:46.594: INFO @metrics_hook: valid age_mae: 6.292 +-0.089 (17639) 2020-02-01 22:21:46.595: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 22:21:48.066: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:21:48.066: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:21:48.067: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 22:21:48.067: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:21:48.207: INFO @evaluate_confidence: Previous accuracy would be: 95.54 2020-02-01 22:21:48.207: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:21:48.267: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.16, 98.23, 98.29, 98.34, 98.39, 98.44, 98.49, 98.55, 98.58, 98.62, 98.68, 98.72, 98.77, 98.81, 98.86, 98.89, 98.93, 98.97, 99.0, 99.04, 99.07, 99.1, 99.13, 99.15, 99.18, 99.2, 99.23, 99.25, 99.28, 99.3, 99.33, 99.36, 99.37, 99.4, 99.41, 99.43, 99.47, 99.49, 99.5, 99.53, 99.54, 99.57, 99.58, 99.6, 99.62, 99.64, 99.65] 2020-02-01 22:21:48.268: INFO @evaluate_confidence: Dropped ratios are: [10.65, 11.09, 11.52, 11.95, 12.38, 12.82, 13.24, 13.64, 14.02, 14.4, 14.82, 15.22, 15.62, 16.05, 16.46, 16.88, 17.31, 17.76, 18.16, 18.59, 19.01, 19.42, 19.85, 20.3, 20.75, 21.21, 21.64, 22.11, 22.56, 23.05, 23.58, 24.11, 24.58, 25.11, 25.7, 26.26, 26.79, 27.32, 27.89, 28.55, 29.17, 29.83, 30.44, 31.13, 31.83, 32.56, 33.29] 2020-02-01 22:21:48.317: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:21:48.317: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:21:48.317: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:21:48.317: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:21:48.456: INFO @evaluate_confidence: Previous accuracy would be: 59.03 2020-02-01 22:21:48.456: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:21:48.471: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.22, 67.86, 68.57, 69.26, 70.03, 70.75, 71.55] 2020-02-01 22:21:48.471: INFO @evaluate_confidence: Dropped ratios are: [41.4, 44.63, 47.82, 50.95, 53.92, 56.75, 59.59] 2020-02-01 22:21:48.479: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:21:48.479: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 22:21:48.479: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.24 2020-02-01 22:21:48.479: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:21:48.584: INFO @evaluate_confidence: Previous accuracy would be: 92.61 2020-02-01 22:21:48.584: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:21:48.593: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.59, 96.69, 96.77, 96.87, 96.93, 97.0, 97.05, 97.11, 97.21, 97.26, 97.34, 97.38, 97.44, 97.53, 97.59, 97.68, 97.78, 97.83, 97.87, 97.91, 97.95, 97.97, 98.0, 98.11, 98.17, 98.2, 98.23, 98.28, 98.29, 98.33, 98.37, 98.42, 98.46, 98.54, 98.57, 98.64, 98.7, 98.74, 98.78, 98.82, 98.83, 98.88] 2020-02-01 22:21:48.593: INFO @evaluate_confidence: Dropped ratios are: [12.86, 13.29, 13.62, 14.01, 14.46, 14.88, 15.18, 15.56, 16.06, 16.38, 16.8, 17.1, 17.48, 17.95, 18.48, 18.94, 19.34, 19.71, 20.14, 20.48, 20.83, 21.32, 21.72, 22.21, 22.69, 23.14, 23.64, 24.25, 24.73, 25.27, 25.74, 26.26, 26.76, 27.39, 28.02, 28.69, 29.42, 30.08, 30.84, 31.54, 32.35, 33.17] 2020-02-01 22:21:48.601: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:21:48.601: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 22:21:48.601: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 22:21:48.601: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:21:48.726: INFO @evaluate_confidence: Previous accuracy would be: 52.24 2020-02-01 22:21:48.727: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 22:21:48.728: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.47, 57.66, 58.48, 58.92] 2020-02-01 22:21:48.728: INFO @evaluate_confidence: Dropped ratios are: [43.89, 48.07, 52.38, 57.02] 2020-02-01 22:21:48.782: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:21:49.525: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:21:49.613: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:21:50.068: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:21:50.146: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:21:50.851: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:21:50.936: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 22:21:50.938: INFO @evaluate_gender-age_model: groups 0 3.462226 1 3.889686 2 5.117559 3 5.548376 4 6.225863 5 6.069998 6 6.063326 7 6.893166 Name: errors, dtype: float64 2020-02-01 22:21:50.939: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:21:51.390: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:21:51.449: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 22:21:51.450: INFO @evaluate_gender-age_model: groups 0 6.114861 1 5.126917 2 5.595881 3 5.723957 4 7.346859 5 5.367239 6 7.970478 7 11.638767 Name: errors, dtype: float64 2020-02-01 22:21:51.638: INFO @decay_lr : LR updated to `4.9571416e-05` 2020-02-01 22:21:51.640: INFO @log_profile : T train: 130.069070 2020-02-01 22:21:51.640: INFO @log_profile : T valid: 5.594130 2020-02-01 22:21:51.640: INFO @log_profile : T read data: 2.829501 2020-02-01 22:21:51.640: INFO @log_profile : T hooks: 10.234976 2020-02-01 22:21:51.640: INFO @main_loop : Epoch 140 done 2020-02-01 22:21:51.640: INFO @main_loop : Training epoch 141 2020-02-01 22:24:11.169: INFO @log_variables: train loss nanmean: 0.686576 2020-02-01 22:24:11.169: INFO @log_variables: train age_loss mean: 5.027192 2020-02-01 22:24:11.170: INFO @log_variables: train gender_loss mean: 0.117338 2020-02-01 22:24:11.170: INFO @log_variables: train age_mae mean: 5.502479 2020-02-01 22:24:11.170: INFO @log_variables: train gender_accuracy mean: 0.953613 2020-02-01 22:24:11.170: INFO @log_variables: train gender_confidence/loss nanmean: 0.051889 2020-02-01 22:24:11.170: INFO @log_variables: train gender_confidence/accuracy mean: 0.853633 2020-02-01 22:24:11.170: INFO @log_variables: train age_confidence/loss mean: 0.070998 2020-02-01 22:24:11.170: INFO @log_variables: train age_confidence/accuracy mean: 0.610026 2020-02-01 22:24:11.170: INFO @log_variables: valid loss nanmean: 0.840703 2020-02-01 22:24:11.170: INFO @log_variables: valid age_loss mean: 5.766325 2020-02-01 22:24:11.170: INFO @log_variables: valid gender_loss mean: 0.209618 2020-02-01 22:24:11.170: INFO @log_variables: valid age_mae mean: 6.246608 2020-02-01 22:24:11.170: INFO @log_variables: valid gender_accuracy mean: 0.923011 2020-02-01 22:24:11.170: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055909 2020-02-01 22:24:11.170: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872895 2020-02-01 22:24:11.170: INFO @log_variables: valid age_confidence/loss mean: 0.070021 2020-02-01 22:24:11.170: INFO @log_variables: valid age_confidence/accuracy mean: 0.561767 2020-02-01 22:24:11.170: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:24:11.177: INFO @metrics_hook: train age_mae: 5.502 +-0.032 (110592) 2020-02-01 22:24:11.184: INFO @metrics_hook: train gender_accuracy: 0.954 +-0.001 (110592) 2020-02-01 22:24:13.892: INFO @metrics_hook: valid age_mae: 6.247 +-0.088 (17639) 2020-02-01 22:24:13.893: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:24:15.538: INFO @decay_lr : LR updated to `4.932356e-05` 2020-02-01 22:24:15.540: INFO @log_profile : T train: 130.549155 2020-02-01 22:24:15.540: INFO @log_profile : T valid: 6.345724 2020-02-01 22:24:15.540: INFO @log_profile : T read data: 1.945430 2020-02-01 22:24:15.540: INFO @log_profile : T hooks: 4.984119 2020-02-01 22:24:15.540: INFO @main_loop : Epoch 141 done 2020-02-01 22:24:15.540: INFO @main_loop : Training epoch 142 2020-02-01 22:26:35.442: INFO @log_variables: train loss nanmean: 0.681582 2020-02-01 22:26:35.443: INFO @log_variables: train age_loss mean: 4.999620 2020-02-01 22:26:35.443: INFO @log_variables: train gender_loss mean: 0.114462 2020-02-01 22:26:35.443: INFO @log_variables: train age_mae mean: 5.474812 2020-02-01 22:26:35.443: INFO @log_variables: train gender_accuracy mean: 0.955867 2020-02-01 22:26:35.443: INFO @log_variables: train gender_confidence/loss nanmean: 0.052074 2020-02-01 22:26:35.443: INFO @log_variables: train gender_confidence/accuracy mean: 0.856667 2020-02-01 22:26:35.443: INFO @log_variables: train age_confidence/loss mean: 0.070941 2020-02-01 22:26:35.443: INFO @log_variables: train age_confidence/accuracy mean: 0.611088 2020-02-01 22:26:35.443: INFO @log_variables: valid loss nanmean: 0.837656 2020-02-01 22:26:35.443: INFO @log_variables: valid age_loss mean: 5.828520 2020-02-01 22:26:35.443: INFO @log_variables: valid gender_loss mean: 0.200650 2020-02-01 22:26:35.443: INFO @log_variables: valid age_mae mean: 6.309770 2020-02-01 22:26:35.443: INFO @log_variables: valid gender_accuracy mean: 0.927717 2020-02-01 22:26:35.443: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055882 2020-02-01 22:26:35.443: INFO @log_variables: valid gender_confidence/accuracy mean: 0.883894 2020-02-01 22:26:35.443: INFO @log_variables: valid age_confidence/loss mean: 0.069499 2020-02-01 22:26:35.444: INFO @log_variables: valid age_confidence/accuracy mean: 0.569590 2020-02-01 22:26:35.444: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:26:35.451: INFO @metrics_hook: train age_mae: 5.475 +-0.032 (110372) 2020-02-01 22:26:35.458: INFO @metrics_hook: train gender_accuracy: 0.956 +-0.001 (110372) 2020-02-01 22:26:38.198: INFO @metrics_hook: valid age_mae: 6.310 +-0.088 (17639) 2020-02-01 22:26:38.199: INFO @metrics_hook: valid gender_accuracy: 0.928 +-0.004 (17639) 2020-02-01 22:26:39.652: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:26:39.653: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:26:39.653: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 22:26:39.653: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:26:39.785: INFO @evaluate_confidence: Previous accuracy would be: 95.59 2020-02-01 22:26:39.785: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:26:39.846: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.11, 98.17, 98.23, 98.28, 98.33, 98.4, 98.45, 98.5, 98.56, 98.61, 98.68, 98.73, 98.78, 98.83, 98.87, 98.91, 98.94, 98.99, 99.01, 99.04, 99.07, 99.1, 99.14, 99.18, 99.21, 99.23, 99.25, 99.28, 99.3, 99.32, 99.36, 99.38, 99.39, 99.41, 99.44, 99.47, 99.49, 99.51, 99.51, 99.53, 99.56, 99.58, 99.6, 99.62, 99.63, 99.64, 99.66] 2020-02-01 22:26:39.846: INFO @evaluate_confidence: Dropped ratios are: [10.65, 11.09, 11.52, 11.92, 12.34, 12.76, 13.16, 13.58, 14.03, 14.41, 14.84, 15.27, 15.67, 16.08, 16.51, 16.93, 17.33, 17.76, 18.2, 18.62, 19.06, 19.48, 19.91, 20.36, 20.78, 21.25, 21.69, 22.16, 22.63, 23.12, 23.59, 24.09, 24.59, 25.1, 25.61, 26.18, 26.74, 27.29, 27.86, 28.42, 29.03, 29.63, 30.26, 30.88, 31.57, 32.35, 33.09] 2020-02-01 22:26:39.895: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:26:39.895: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:26:39.895: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:26:39.896: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:26:40.032: INFO @evaluate_confidence: Previous accuracy would be: 58.76 2020-02-01 22:26:40.032: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:26:40.047: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.03, 67.63, 68.27, 68.91, 69.72, 70.39, 71.2] 2020-02-01 22:26:40.047: INFO @evaluate_confidence: Dropped ratios are: [41.66, 44.8, 47.94, 51.06, 54.06, 56.91, 59.76] 2020-02-01 22:26:40.055: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:26:40.055: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.87 +- 0.21 2020-02-01 22:26:40.055: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.24 2020-02-01 22:26:40.055: INFO @evaluate_confidence: Average confidence of all samples 0.84 +- 0.23 2020-02-01 22:26:40.155: INFO @evaluate_confidence: Previous accuracy would be: 92.77 2020-02-01 22:26:40.155: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 22:26:40.163: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.43, 96.54, 96.62, 96.67, 96.79, 96.89, 96.94, 97.05, 97.11, 97.21, 97.3, 97.35, 97.42, 97.46, 97.51, 97.6, 97.68, 97.71, 97.75, 97.8, 97.9, 97.94, 97.99, 98.05, 98.13, 98.2, 98.25, 98.28, 98.32, 98.41, 98.48, 98.53, 98.59, 98.65, 98.69, 98.81, 98.86, 98.89] 2020-02-01 22:26:40.163: INFO @evaluate_confidence: Dropped ratios are: [12.44, 12.81, 13.18, 13.5, 13.92, 14.34, 14.64, 15.04, 15.36, 15.68, 16.04, 16.44, 16.8, 17.13, 17.58, 17.98, 18.46, 18.83, 19.26, 19.68, 20.21, 20.66, 21.15, 21.61, 22.14, 22.67, 23.2, 23.71, 24.28, 24.88, 25.48, 26.14, 26.88, 27.67, 28.4, 29.24, 30.06, 30.89] 2020-02-01 22:26:40.171: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:26:40.171: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.10 2020-02-01 22:26:40.171: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 22:26:40.171: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 22:26:40.297: INFO @evaluate_confidence: Previous accuracy would be: 51.85 2020-02-01 22:26:40.297: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 22:26:40.299: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.98, 58.26, 58.57, 59.05, 59.65] 2020-02-01 22:26:40.299: INFO @evaluate_confidence: Dropped ratios are: [44.86, 49.69, 54.24, 58.74, 63.36] 2020-02-01 22:26:40.350: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:26:41.036: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:26:41.119: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:26:41.573: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:26:41.830: INFO @decay_lr : LR updated to `4.907694e-05` 2020-02-01 22:26:42.200: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 22:26:42.205: INFO @log_profile : T train: 130.099326 2020-02-01 22:26:42.205: INFO @log_profile : T valid: 6.306589 2020-02-01 22:26:42.205: INFO @log_profile : T read data: 2.813702 2020-02-01 22:26:42.205: INFO @log_profile : T hooks: 7.369796 2020-02-01 22:26:42.205: INFO @main_loop : Epoch 142 done 2020-02-01 22:26:42.205: INFO @main_loop : Training epoch 143 2020-02-01 22:28:56.503: INFO @log_variables: train loss nanmean: 0.678642 2020-02-01 22:28:56.503: INFO @log_variables: train age_loss mean: 4.981286 2020-02-01 22:28:56.503: INFO @log_variables: train gender_loss mean: 0.114116 2020-02-01 22:28:56.503: INFO @log_variables: train age_mae mean: 5.456266 2020-02-01 22:28:56.503: INFO @log_variables: train gender_accuracy mean: 0.954653 2020-02-01 22:28:56.503: INFO @log_variables: train gender_confidence/loss nanmean: 0.050907 2020-02-01 22:28:56.503: INFO @log_variables: train gender_confidence/accuracy mean: 0.859584 2020-02-01 22:28:56.503: INFO @log_variables: train age_confidence/loss mean: 0.071149 2020-02-01 22:28:56.504: INFO @log_variables: train age_confidence/accuracy mean: 0.609185 2020-02-01 22:28:56.504: INFO @log_variables: valid loss nanmean: 0.840279 2020-02-01 22:28:56.504: INFO @log_variables: valid age_loss mean: 5.800248 2020-02-01 22:28:56.504: INFO @log_variables: valid gender_loss mean: 0.207207 2020-02-01 22:28:56.504: INFO @log_variables: valid age_mae mean: 6.280055 2020-02-01 22:28:56.504: INFO @log_variables: valid gender_accuracy mean: 0.922898 2020-02-01 22:28:56.504: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054362 2020-02-01 22:28:56.504: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876580 2020-02-01 22:28:56.504: INFO @log_variables: valid age_confidence/loss mean: 0.070252 2020-02-01 22:28:56.504: INFO @log_variables: valid age_confidence/accuracy mean: 0.567492 2020-02-01 22:28:56.504: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:28:56.511: INFO @metrics_hook: train age_mae: 5.456 +-0.031 (110372) 2020-02-01 22:28:56.519: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110372) 2020-02-01 22:28:59.314: INFO @metrics_hook: valid age_mae: 6.280 +-0.089 (17639) 2020-02-01 22:28:59.315: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:29:00.997: INFO @decay_lr : LR updated to `4.8831556e-05` 2020-02-01 22:29:00.999: INFO @log_profile : T train: 125.207583 2020-02-01 22:29:00.999: INFO @log_profile : T valid: 5.585745 2020-02-01 22:29:00.999: INFO @log_profile : T read data: 2.836599 2020-02-01 22:29:00.999: INFO @log_profile : T hooks: 5.086895 2020-02-01 22:29:00.999: INFO @main_loop : Epoch 143 done 2020-02-01 22:29:00.999: INFO @main_loop : Training epoch 144 2020-02-01 22:31:20.196: INFO @log_variables: train loss nanmean: 0.682256 2020-02-01 22:31:20.197: INFO @log_variables: train age_loss mean: 5.006987 2020-02-01 22:31:20.197: INFO @log_variables: train gender_loss mean: 0.114064 2020-02-01 22:31:20.197: INFO @log_variables: train age_mae mean: 5.482191 2020-02-01 22:31:20.197: INFO @log_variables: train gender_accuracy mean: 0.955458 2020-02-01 22:31:20.197: INFO @log_variables: train gender_confidence/loss nanmean: 0.052353 2020-02-01 22:31:20.197: INFO @log_variables: train gender_confidence/accuracy mean: 0.854971 2020-02-01 22:31:20.197: INFO @log_variables: train age_confidence/loss mean: 0.071028 2020-02-01 22:31:20.197: INFO @log_variables: train age_confidence/accuracy mean: 0.608959 2020-02-01 22:31:20.197: INFO @log_variables: valid loss nanmean: 0.860950 2020-02-01 22:31:20.197: INFO @log_variables: valid age_loss mean: 5.794741 2020-02-01 22:31:20.197: INFO @log_variables: valid gender_loss mean: 0.227661 2020-02-01 22:31:20.197: INFO @log_variables: valid age_mae mean: 6.275256 2020-02-01 22:31:20.197: INFO @log_variables: valid gender_accuracy mean: 0.917286 2020-02-01 22:31:20.197: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057086 2020-02-01 22:31:20.197: INFO @log_variables: valid gender_confidence/accuracy mean: 0.862521 2020-02-01 22:31:20.197: INFO @log_variables: valid age_confidence/loss mean: 0.070105 2020-02-01 22:31:20.197: INFO @log_variables: valid age_confidence/accuracy mean: 0.552922 2020-02-01 22:31:20.197: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:31:20.204: INFO @metrics_hook: train age_mae: 5.482 +-0.032 (110592) 2020-02-01 22:31:20.211: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110592) 2020-02-01 22:31:22.911: INFO @metrics_hook: valid age_mae: 6.275 +-0.088 (17639) 2020-02-01 22:31:22.912: INFO @metrics_hook: valid gender_accuracy: 0.917 +-0.004 (17639) 2020-02-01 22:31:24.381: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:31:24.381: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.82 +- 0.24 2020-02-01 22:31:24.381: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.35 +- 0.21 2020-02-01 22:31:24.382: INFO @evaluate_confidence: Average confidence of all samples 0.80 +- 0.26 2020-02-01 22:31:24.509: INFO @evaluate_confidence: Previous accuracy would be: 95.55 2020-02-01 22:31:24.510: INFO @evaluate_confidence: Possible optimal thresholds are: [0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81] 2020-02-01 22:31:24.570: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.13, 98.21, 98.27, 98.33, 98.4, 98.45, 98.5, 98.56, 98.61, 98.66, 98.71, 98.75, 98.8, 98.84, 98.87, 98.91, 98.94, 98.97, 98.99, 99.02, 99.06, 99.08, 99.12, 99.14, 99.16, 99.2, 99.23, 99.26, 99.29, 99.3, 99.33, 99.35, 99.36, 99.38, 99.4, 99.43, 99.46, 99.48, 99.5, 99.52, 99.54, 99.56, 99.58, 99.59, 99.61, 99.62, 99.64] 2020-02-01 22:31:24.570: INFO @evaluate_confidence: Dropped ratios are: [10.86, 11.29, 11.71, 12.14, 12.57, 13.0, 13.41, 13.85, 14.27, 14.7, 15.11, 15.52, 15.93, 16.35, 16.76, 17.16, 17.59, 18.01, 18.42, 18.82, 19.22, 19.66, 20.1, 20.51, 20.94, 21.38, 21.84, 22.34, 22.83, 23.35, 23.77, 24.25, 24.73, 25.26, 25.75, 26.28, 26.8, 27.37, 27.9, 28.5, 29.09, 29.73, 30.38, 31.04, 31.73, 32.44, 33.19] 2020-02-01 22:31:24.619: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:31:24.619: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:31:24.619: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:31:24.619: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:31:24.758: INFO @evaluate_confidence: Previous accuracy would be: 58.82 2020-02-01 22:31:24.758: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:31:24.773: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.86, 67.55, 68.22, 68.96, 69.68, 70.48, 71.25] 2020-02-01 22:31:24.773: INFO @evaluate_confidence: Dropped ratios are: [41.51, 44.72, 47.84, 50.98, 54.01, 56.93, 59.71] 2020-02-01 22:31:24.780: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:31:24.780: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 22:31:24.781: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.24 2020-02-01 22:31:24.781: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:31:24.883: INFO @evaluate_confidence: Previous accuracy would be: 91.73 2020-02-01 22:31:24.883: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:31:24.892: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.93, 96.01, 96.1, 96.19, 96.25, 96.35, 96.44, 96.52, 96.63, 96.74, 96.8, 96.91, 97.0, 97.05, 97.09, 97.13, 97.18, 97.24, 97.32, 97.41, 97.44, 97.53, 97.59, 97.66, 97.68, 97.71, 97.8, 97.86, 97.93, 97.99, 98.03, 98.1, 98.2, 98.29, 98.36, 98.46, 98.55, 98.58, 98.65, 98.72] 2020-02-01 22:31:24.892: INFO @evaluate_confidence: Dropped ratios are: [14.0, 14.43, 14.78, 15.28, 15.69, 16.08, 16.53, 17.01, 17.39, 17.82, 18.32, 18.67, 19.12, 19.57, 20.07, 20.51, 20.85, 21.21, 21.69, 22.13, 22.67, 23.23, 23.71, 24.24, 24.64, 25.15, 25.65, 26.19, 26.86, 27.36, 27.94, 28.52, 29.16, 29.75, 30.34, 30.98, 31.8, 32.37, 33.11, 33.94] 2020-02-01 22:31:24.899: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:31:24.900: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 22:31:24.900: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.09 2020-02-01 22:31:24.900: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:31:25.025: INFO @evaluate_confidence: Previous accuracy would be: 52.42 2020-02-01 22:31:25.025: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 22:31:25.026: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.53, 57.58, 58.19, 58.59] 2020-02-01 22:31:25.026: INFO @evaluate_confidence: Dropped ratios are: [44.8, 49.17, 53.97, 58.76] 2020-02-01 22:31:25.078: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:31:25.784: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:31:25.871: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:31:26.350: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:31:26.616: INFO @decay_lr : LR updated to `4.8587397e-05` 2020-02-01 22:31:26.618: INFO @log_profile : T train: 130.227486 2020-02-01 22:31:26.618: INFO @log_profile : T valid: 6.358012 2020-02-01 22:31:26.618: INFO @log_profile : T read data: 1.942291 2020-02-01 22:31:26.618: INFO @log_profile : T hooks: 7.015634 2020-02-01 22:31:26.618: INFO @main_loop : Epoch 144 done 2020-02-01 22:31:26.618: INFO @main_loop : Training epoch 145 2020-02-01 22:33:46.185: INFO @log_variables: train loss nanmean: 0.682018 2020-02-01 22:33:46.185: INFO @log_variables: train age_loss mean: 5.001939 2020-02-01 22:33:46.185: INFO @log_variables: train gender_loss mean: 0.114613 2020-02-01 22:33:46.185: INFO @log_variables: train age_mae mean: 5.476915 2020-02-01 22:33:46.185: INFO @log_variables: train gender_accuracy mean: 0.955088 2020-02-01 22:33:46.186: INFO @log_variables: train gender_confidence/loss nanmean: 0.052075 2020-02-01 22:33:46.186: INFO @log_variables: train gender_confidence/accuracy mean: 0.855905 2020-02-01 22:33:46.186: INFO @log_variables: train age_confidence/loss mean: 0.071028 2020-02-01 22:33:46.186: INFO @log_variables: train age_confidence/accuracy mean: 0.606821 2020-02-01 22:33:46.186: INFO @log_variables: valid loss nanmean: 0.824244 2020-02-01 22:33:46.186: INFO @log_variables: valid age_loss mean: 5.731740 2020-02-01 22:33:46.186: INFO @log_variables: valid gender_loss mean: 0.195628 2020-02-01 22:33:46.186: INFO @log_variables: valid age_mae mean: 6.211803 2020-02-01 22:33:46.186: INFO @log_variables: valid gender_accuracy mean: 0.926640 2020-02-01 22:33:46.186: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054685 2020-02-01 22:33:46.186: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871591 2020-02-01 22:33:46.186: INFO @log_variables: valid age_confidence/loss mean: 0.070648 2020-02-01 22:33:46.186: INFO @log_variables: valid age_confidence/accuracy mean: 0.562163 2020-02-01 22:33:46.186: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:33:46.193: INFO @metrics_hook: train age_mae: 5.477 +-0.032 (110372) 2020-02-01 22:33:46.200: INFO @metrics_hook: train gender_accuracy: 0.955 +-0.001 (110372) 2020-02-01 22:33:48.927: INFO @metrics_hook: valid age_mae: 6.212 +-0.089 (17639) 2020-02-01 22:33:48.928: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 22:33:50.599: INFO @decay_lr : LR updated to `4.834446e-05` 2020-02-01 22:33:50.600: INFO @log_profile : T train: 130.407184 2020-02-01 22:33:50.600: INFO @log_profile : T valid: 5.650144 2020-02-01 22:33:50.600: INFO @log_profile : T read data: 2.837444 2020-02-01 22:33:50.600: INFO @log_profile : T hooks: 5.013162 2020-02-01 22:33:50.600: INFO @main_loop : Epoch 145 done 2020-02-01 22:33:50.600: INFO @main_loop : Training epoch 146 2020-02-01 22:36:08.266: INFO @log_variables: train loss nanmean: 0.677057 2020-02-01 22:36:08.266: INFO @log_variables: train age_loss mean: 4.977450 2020-02-01 22:36:08.267: INFO @log_variables: train gender_loss mean: 0.112653 2020-02-01 22:36:08.267: INFO @log_variables: train age_mae mean: 5.452481 2020-02-01 22:36:08.267: INFO @log_variables: train gender_accuracy mean: 0.956166 2020-02-01 22:36:08.267: INFO @log_variables: train gender_confidence/loss nanmean: 0.050989 2020-02-01 22:36:08.267: INFO @log_variables: train gender_confidence/accuracy mean: 0.857482 2020-02-01 22:36:08.267: INFO @log_variables: train age_confidence/loss mean: 0.071161 2020-02-01 22:36:08.267: INFO @log_variables: train age_confidence/accuracy mean: 0.608062 2020-02-01 22:36:08.267: INFO @log_variables: valid loss nanmean: 0.841210 2020-02-01 22:36:08.267: INFO @log_variables: valid age_loss mean: 5.849291 2020-02-01 22:36:08.267: INFO @log_variables: valid gender_loss mean: 0.202472 2020-02-01 22:36:08.267: INFO @log_variables: valid age_mae mean: 6.328187 2020-02-01 22:36:08.267: INFO @log_variables: valid gender_accuracy mean: 0.923408 2020-02-01 22:36:08.267: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054444 2020-02-01 22:36:08.267: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867906 2020-02-01 22:36:08.267: INFO @log_variables: valid age_confidence/loss mean: 0.070947 2020-02-01 22:36:08.267: INFO @log_variables: valid age_confidence/accuracy mean: 0.555927 2020-02-01 22:36:08.267: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:36:08.275: INFO @metrics_hook: train age_mae: 5.452 +-0.031 (110372) 2020-02-01 22:36:08.282: INFO @metrics_hook: train gender_accuracy: 0.956 +-0.001 (110372) 2020-02-01 22:36:11.032: INFO @metrics_hook: valid age_mae: 6.328 +-0.092 (17639) 2020-02-01 22:36:11.033: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 22:36:12.528: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:36:12.528: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:36:12.528: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:36:12.529: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:36:12.663: INFO @evaluate_confidence: Previous accuracy would be: 95.62 2020-02-01 22:36:12.663: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:36:12.727: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.14, 98.2, 98.27, 98.33, 98.38, 98.43, 98.5, 98.54, 98.6, 98.64, 98.69, 98.73, 98.78, 98.81, 98.84, 98.88, 98.93, 98.97, 99.01, 99.06, 99.08, 99.12, 99.15, 99.18, 99.19, 99.22, 99.24, 99.27, 99.3, 99.33, 99.36, 99.37, 99.4, 99.43, 99.46, 99.48, 99.5, 99.52, 99.54, 99.57, 99.59, 99.61, 99.63, 99.64, 99.66, 99.68, 99.7, 99.71, 99.73] 2020-02-01 22:36:12.727: INFO @evaluate_confidence: Dropped ratios are: [10.29, 10.69, 11.12, 11.52, 11.95, 12.37, 12.78, 13.21, 13.61, 14.03, 14.44, 14.81, 15.22, 15.61, 16.01, 16.45, 16.85, 17.28, 17.7, 18.12, 18.54, 18.95, 19.35, 19.79, 20.21, 20.63, 21.05, 21.53, 21.96, 22.42, 22.89, 23.38, 23.85, 24.33, 24.84, 25.36, 25.85, 26.36, 26.9, 27.49, 28.04, 28.64, 29.28, 29.88, 30.53, 31.2, 31.93, 32.66, 33.46] 2020-02-01 22:36:12.777: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:36:12.777: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:36:12.777: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:36:12.777: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:36:12.918: INFO @evaluate_confidence: Previous accuracy would be: 59.03 2020-02-01 22:36:12.918: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:36:12.933: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.84, 67.47, 68.19, 69.05, 69.73, 70.63, 71.52] 2020-02-01 22:36:12.934: INFO @evaluate_confidence: Dropped ratios are: [41.1, 44.27, 47.4, 50.59, 53.51, 56.53, 59.29] 2020-02-01 22:36:12.941: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:36:12.941: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 22:36:12.942: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.24 2020-02-01 22:36:12.942: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:36:13.051: INFO @evaluate_confidence: Previous accuracy would be: 92.34 2020-02-01 22:36:13.051: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:36:13.061: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.24, 96.36, 96.46, 96.57, 96.68, 96.75, 96.79, 96.92, 97.03, 97.08, 97.15, 97.26, 97.32, 97.38, 97.44, 97.55, 97.61, 97.67, 97.74, 97.84, 97.89, 97.99, 98.05, 98.08, 98.16, 98.26, 98.31, 98.41, 98.43, 98.47, 98.55, 98.59, 98.63, 98.64, 98.68, 98.72, 98.76, 98.81, 98.88, 98.94, 98.99, 99.04] 2020-02-01 22:36:13.061: INFO @evaluate_confidence: Dropped ratios are: [12.89, 13.33, 13.86, 14.28, 14.69, 15.1, 15.44, 15.87, 16.34, 16.71, 17.13, 17.67, 18.0, 18.45, 18.85, 19.24, 19.64, 20.09, 20.54, 21.03, 21.48, 21.96, 22.39, 22.81, 23.22, 23.69, 24.33, 24.87, 25.4, 25.86, 26.41, 26.99, 27.63, 28.26, 28.87, 29.45, 29.95, 30.65, 31.47, 32.22, 33.06, 33.9] 2020-02-01 22:36:13.068: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:36:13.069: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 22:36:13.069: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 22:36:13.069: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.11 2020-02-01 22:36:13.198: INFO @evaluate_confidence: Previous accuracy would be: 53.11 2020-02-01 22:36:13.198: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 22:36:13.199: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.07, 58.42, 58.66] 2020-02-01 22:36:13.199: INFO @evaluate_confidence: Dropped ratios are: [46.07, 50.59, 54.91] 2020-02-01 22:36:13.252: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:36:13.939: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:36:14.023: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:36:14.492: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:36:14.762: INFO @decay_lr : LR updated to `4.810274e-05` 2020-02-01 22:36:14.764: INFO @log_profile : T train: 129.522343 2020-02-01 22:36:14.764: INFO @log_profile : T valid: 5.619751 2020-02-01 22:36:14.764: INFO @log_profile : T read data: 1.849880 2020-02-01 22:36:14.764: INFO @log_profile : T hooks: 7.095453 2020-02-01 22:36:14.764: INFO @main_loop : Epoch 146 done 2020-02-01 22:36:14.764: INFO @main_loop : Training epoch 147 2020-02-01 22:38:26.029: INFO @log_variables: train loss nanmean: 0.673473 2020-02-01 22:38:26.029: INFO @log_variables: train age_loss mean: 4.944466 2020-02-01 22:38:26.029: INFO @log_variables: train gender_loss mean: 0.112065 2020-02-01 22:38:26.029: INFO @log_variables: train age_mae mean: 5.419357 2020-02-01 22:38:26.029: INFO @log_variables: train gender_accuracy mean: 0.955684 2020-02-01 22:38:26.029: INFO @log_variables: train gender_confidence/loss nanmean: 0.050941 2020-02-01 22:38:26.029: INFO @log_variables: train gender_confidence/accuracy mean: 0.858588 2020-02-01 22:38:26.029: INFO @log_variables: train age_confidence/loss mean: 0.071158 2020-02-01 22:38:26.030: INFO @log_variables: train age_confidence/accuracy mean: 0.613182 2020-02-01 22:38:26.030: INFO @log_variables: valid loss nanmean: 0.836373 2020-02-01 22:38:26.030: INFO @log_variables: valid age_loss mean: 5.723271 2020-02-01 22:38:26.030: INFO @log_variables: valid gender_loss mean: 0.207502 2020-02-01 22:38:26.030: INFO @log_variables: valid age_mae mean: 6.203037 2020-02-01 22:38:26.030: INFO @log_variables: valid gender_accuracy mean: 0.926923 2020-02-01 22:38:26.030: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057041 2020-02-01 22:38:26.030: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877317 2020-02-01 22:38:26.030: INFO @log_variables: valid age_confidence/loss mean: 0.070397 2020-02-01 22:38:26.030: INFO @log_variables: valid age_confidence/accuracy mean: 0.568740 2020-02-01 22:38:26.030: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:38:26.038: INFO @metrics_hook: train age_mae: 5.419 +-0.031 (110592) 2020-02-01 22:38:26.045: INFO @metrics_hook: train gender_accuracy: 0.956 +-0.001 (110592) 2020-02-01 22:38:28.809: INFO @metrics_hook: valid age_mae: 6.203 +-0.088 (17639) 2020-02-01 22:38:28.810: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 22:38:30.454: INFO @decay_lr : LR updated to `4.7862224e-05` 2020-02-01 22:38:30.455: INFO @log_profile : T train: 122.126351 2020-02-01 22:38:30.455: INFO @log_profile : T valid: 5.473102 2020-02-01 22:38:30.455: INFO @log_profile : T read data: 2.956950 2020-02-01 22:38:30.455: INFO @log_profile : T hooks: 5.057614 2020-02-01 22:38:30.455: INFO @main_loop : Epoch 147 done 2020-02-01 22:38:30.455: INFO @main_loop : Training epoch 148 2020-02-01 22:40:40.978: INFO @log_variables: train loss nanmean: 0.671157 2020-02-01 22:40:40.978: INFO @log_variables: train age_loss mean: 4.950973 2020-02-01 22:40:40.978: INFO @log_variables: train gender_loss mean: 0.108735 2020-02-01 22:40:40.978: INFO @log_variables: train age_mae mean: 5.425985 2020-02-01 22:40:40.978: INFO @log_variables: train gender_accuracy mean: 0.957417 2020-02-01 22:40:40.978: INFO @log_variables: train gender_confidence/loss nanmean: 0.050940 2020-02-01 22:40:40.978: INFO @log_variables: train gender_confidence/accuracy mean: 0.861097 2020-02-01 22:40:40.978: INFO @log_variables: train age_confidence/loss mean: 0.071279 2020-02-01 22:40:40.978: INFO @log_variables: train age_confidence/accuracy mean: 0.608316 2020-02-01 22:40:40.978: INFO @log_variables: valid loss nanmean: 0.839519 2020-02-01 22:40:40.978: INFO @log_variables: valid age_loss mean: 5.768909 2020-02-01 22:40:40.978: INFO @log_variables: valid gender_loss mean: 0.210139 2020-02-01 22:40:40.978: INFO @log_variables: valid age_mae mean: 6.249574 2020-02-01 22:40:40.978: INFO @log_variables: valid gender_accuracy mean: 0.923919 2020-02-01 22:40:40.978: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054088 2020-02-01 22:40:40.978: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874256 2020-02-01 22:40:40.979: INFO @log_variables: valid age_confidence/loss mean: 0.069949 2020-02-01 22:40:40.979: INFO @log_variables: valid age_confidence/accuracy mean: 0.569477 2020-02-01 22:40:40.979: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:40:40.986: INFO @metrics_hook: train age_mae: 5.426 +-0.031 (110372) 2020-02-01 22:40:40.993: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110372) 2020-02-01 22:40:43.728: INFO @metrics_hook: valid age_mae: 6.250 +-0.089 (17639) 2020-02-01 22:40:43.730: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 22:40:45.173: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:40:45.173: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:40:45.173: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:40:45.174: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:40:45.302: INFO @evaluate_confidence: Previous accuracy would be: 95.74 2020-02-01 22:40:45.302: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:40:45.364: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.22, 98.28, 98.32, 98.38, 98.44, 98.48, 98.52, 98.58, 98.63, 98.69, 98.74, 98.78, 98.83, 98.88, 98.92, 98.96, 98.99, 99.02, 99.06, 99.09, 99.11, 99.14, 99.18, 99.21, 99.22, 99.25, 99.29, 99.31, 99.34, 99.37, 99.39, 99.42, 99.44, 99.46, 99.48, 99.5, 99.52, 99.53, 99.55, 99.57, 99.58, 99.61, 99.63, 99.64, 99.66, 99.67, 99.68, 99.69, 99.7] 2020-02-01 22:40:45.364: INFO @evaluate_confidence: Dropped ratios are: [10.03, 10.46, 10.85, 11.25, 11.64, 12.05, 12.45, 12.83, 13.23, 13.65, 14.06, 14.46, 14.85, 15.26, 15.67, 16.07, 16.45, 16.87, 17.31, 17.72, 18.1, 18.54, 18.97, 19.4, 19.79, 20.26, 20.72, 21.15, 21.59, 22.01, 22.46, 22.95, 23.46, 23.95, 24.43, 24.96, 25.48, 25.96, 26.51, 27.08, 27.66, 28.22, 28.82, 29.44, 30.08, 30.75, 31.47, 32.19, 32.98] 2020-02-01 22:40:45.414: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:40:45.414: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:40:45.414: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:40:45.414: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:40:45.551: INFO @evaluate_confidence: Previous accuracy would be: 59.15 2020-02-01 22:40:45.551: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:40:45.566: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [66.97, 67.63, 68.37, 69.16, 69.83, 70.6, 71.37] 2020-02-01 22:40:45.566: INFO @evaluate_confidence: Dropped ratios are: [41.12, 44.27, 47.44, 50.54, 53.6, 56.43, 59.24] 2020-02-01 22:40:45.573: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:40:45.574: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 22:40:45.574: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.25 2020-02-01 22:40:45.574: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 22:40:45.675: INFO @evaluate_confidence: Previous accuracy would be: 92.39 2020-02-01 22:40:45.676: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 22:40:45.684: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.27, 96.38, 96.46, 96.53, 96.66, 96.72, 96.8, 96.85, 96.93, 97.01, 97.06, 97.18, 97.24, 97.3, 97.36, 97.42, 97.47, 97.59, 97.61, 97.71, 97.77, 97.83, 97.86, 97.91, 97.98, 98.02, 98.11, 98.15, 98.2, 98.3, 98.36, 98.42, 98.46, 98.53, 98.6, 98.65, 98.7, 98.75, 98.8, 98.8, 98.88] 2020-02-01 22:40:45.684: INFO @evaluate_confidence: Dropped ratios are: [12.89, 13.3, 13.59, 14.0, 14.47, 14.88, 15.27, 15.53, 15.94, 16.44, 16.78, 17.18, 17.7, 18.09, 18.47, 18.85, 19.2, 19.63, 20.06, 20.57, 21.03, 21.46, 21.85, 22.2, 22.64, 23.14, 23.65, 24.1, 24.58, 25.14, 25.7, 26.25, 26.86, 27.48, 28.09, 28.75, 29.42, 30.04, 30.75, 31.5, 32.34] 2020-02-01 22:40:45.692: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:40:45.692: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 22:40:45.692: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 22:40:45.692: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 22:40:45.817: INFO @evaluate_confidence: Previous accuracy would be: 52.54 2020-02-01 22:40:45.818: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 22:40:45.819: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.68, 58.95, 59.37] 2020-02-01 22:40:45.819: INFO @evaluate_confidence: Dropped ratios are: [45.31, 50.17, 54.99] 2020-02-01 22:40:45.870: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:40:46.556: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:40:46.641: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:40:47.095: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:40:47.352: INFO @decay_lr : LR updated to `4.7622914e-05` 2020-02-01 22:40:47.353: INFO @log_profile : T train: 121.449462 2020-02-01 22:40:47.353: INFO @log_profile : T valid: 5.492215 2020-02-01 22:40:47.353: INFO @log_profile : T read data: 2.899337 2020-02-01 22:40:47.353: INFO @log_profile : T hooks: 6.978711 2020-02-01 22:40:47.353: INFO @main_loop : Epoch 148 done 2020-02-01 22:40:47.353: INFO @main_loop : Training epoch 149 2020-02-01 22:42:57.329: INFO @log_variables: train loss nanmean: 0.673342 2020-02-01 22:42:57.330: INFO @log_variables: train age_loss mean: 4.957609 2020-02-01 22:42:57.330: INFO @log_variables: train gender_loss mean: 0.110003 2020-02-01 22:42:57.330: INFO @log_variables: train age_mae mean: 5.432227 2020-02-01 22:42:57.330: INFO @log_variables: train gender_accuracy mean: 0.957058 2020-02-01 22:42:57.330: INFO @log_variables: train gender_confidence/loss nanmean: 0.051365 2020-02-01 22:42:57.330: INFO @log_variables: train gender_confidence/accuracy mean: 0.856680 2020-02-01 22:42:57.330: INFO @log_variables: train age_confidence/loss mean: 0.071283 2020-02-01 22:42:57.330: INFO @log_variables: train age_confidence/accuracy mean: 0.610939 2020-02-01 22:42:57.330: INFO @log_variables: valid loss nanmean: 0.839899 2020-02-01 22:42:57.330: INFO @log_variables: valid age_loss mean: 5.822501 2020-02-01 22:42:57.330: INFO @log_variables: valid gender_loss mean: 0.204439 2020-02-01 22:42:57.330: INFO @log_variables: valid age_mae mean: 6.301988 2020-02-01 22:42:57.330: INFO @log_variables: valid gender_accuracy mean: 0.923975 2020-02-01 22:42:57.330: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054806 2020-02-01 22:42:57.330: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866602 2020-02-01 22:42:57.330: INFO @log_variables: valid age_confidence/loss mean: 0.069921 2020-02-01 22:42:57.330: INFO @log_variables: valid age_confidence/accuracy mean: 0.566132 2020-02-01 22:42:57.330: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:42:57.338: INFO @metrics_hook: train age_mae: 5.432 +-0.031 (110592) 2020-02-01 22:42:57.345: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110592) 2020-02-01 22:43:00.127: INFO @metrics_hook: valid age_mae: 6.302 +-0.089 (17639) 2020-02-01 22:43:00.128: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 22:43:01.828: INFO @decay_lr : LR updated to `4.73848e-05` 2020-02-01 22:43:01.829: INFO @log_profile : T train: 121.916606 2020-02-01 22:43:01.829: INFO @log_profile : T valid: 5.460267 2020-02-01 22:43:01.829: INFO @log_profile : T read data: 1.894725 2020-02-01 22:43:01.829: INFO @log_profile : T hooks: 5.127199 2020-02-01 22:43:01.829: INFO @main_loop : Epoch 149 done 2020-02-01 22:43:01.829: INFO @main_loop : Training epoch 150 2020-02-01 22:45:14.237: INFO @log_variables: train loss nanmean: 0.672279 2020-02-01 22:45:14.238: INFO @log_variables: train age_loss mean: 4.936447 2020-02-01 22:45:14.238: INFO @log_variables: train gender_loss mean: 0.111395 2020-02-01 22:45:14.238: INFO @log_variables: train age_mae mean: 5.411273 2020-02-01 22:45:14.238: INFO @log_variables: train gender_accuracy mean: 0.955985 2020-02-01 22:45:14.238: INFO @log_variables: train gender_confidence/loss nanmean: 0.050730 2020-02-01 22:45:14.238: INFO @log_variables: train gender_confidence/accuracy mean: 0.858497 2020-02-01 22:45:14.238: INFO @log_variables: train age_confidence/loss mean: 0.071513 2020-02-01 22:45:14.238: INFO @log_variables: train age_confidence/accuracy mean: 0.609104 2020-02-01 22:45:14.238: INFO @log_variables: valid loss nanmean: 0.864744 2020-02-01 22:45:14.238: INFO @log_variables: valid age_loss mean: 5.946143 2020-02-01 22:45:14.238: INFO @log_variables: valid gender_loss mean: 0.219110 2020-02-01 22:45:14.238: INFO @log_variables: valid age_mae mean: 6.427029 2020-02-01 22:45:14.238: INFO @log_variables: valid gender_accuracy mean: 0.920687 2020-02-01 22:45:14.238: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055570 2020-02-01 22:45:14.238: INFO @log_variables: valid gender_confidence/accuracy mean: 0.859743 2020-02-01 22:45:14.238: INFO @log_variables: valid age_confidence/loss mean: 0.069424 2020-02-01 22:45:14.238: INFO @log_variables: valid age_confidence/accuracy mean: 0.562560 2020-02-01 22:45:14.239: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:45:14.247: INFO @metrics_hook: train age_mae: 5.411 +-0.031 (110372) 2020-02-01 22:45:14.254: INFO @metrics_hook: train gender_accuracy: 0.956 +-0.001 (110372) 2020-02-01 22:45:16.955: INFO @metrics_hook: valid age_mae: 6.427 +-0.091 (17639) 2020-02-01 22:45:16.957: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 22:45:18.391: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:45:18.391: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:45:18.391: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:45:18.392: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:45:18.519: INFO @evaluate_confidence: Previous accuracy would be: 95.60 2020-02-01 22:45:18.520: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:45:18.581: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.2, 98.26, 98.32, 98.37, 98.42, 98.49, 98.54, 98.59, 98.64, 98.69, 98.73, 98.77, 98.82, 98.86, 98.89, 98.93, 98.96, 98.99, 99.03, 99.07, 99.12, 99.15, 99.19, 99.22, 99.24, 99.26, 99.3, 99.31, 99.33, 99.35, 99.37, 99.39, 99.42, 99.45, 99.47, 99.49, 99.5, 99.52, 99.55, 99.56, 99.58, 99.59, 99.61, 99.63, 99.65, 99.66, 99.67, 99.69, 99.69] 2020-02-01 22:45:18.581: INFO @evaluate_confidence: Dropped ratios are: [10.39, 10.77, 11.2, 11.59, 12.01, 12.44, 12.85, 13.27, 13.66, 14.09, 14.5, 14.86, 15.26, 15.65, 16.03, 16.43, 16.83, 17.2, 17.57, 18.0, 18.4, 18.82, 19.24, 19.63, 20.05, 20.46, 20.87, 21.3, 21.72, 22.18, 22.63, 23.09, 23.5, 23.97, 24.45, 24.96, 25.47, 25.96, 26.5, 27.07, 27.61, 28.21, 28.83, 29.47, 30.1, 30.75, 31.46, 32.2, 32.93] 2020-02-01 22:45:18.629: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:45:18.629: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.13 2020-02-01 22:45:18.630: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:45:18.630: INFO @evaluate_confidence: Average confidence of all samples 0.54 +- 0.13 2020-02-01 22:45:18.766: INFO @evaluate_confidence: Previous accuracy would be: 59.37 2020-02-01 22:45:18.766: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:45:18.781: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.14, 67.82, 68.46, 69.24, 69.99, 70.79, 71.65] 2020-02-01 22:45:18.781: INFO @evaluate_confidence: Dropped ratios are: [40.84, 44.03, 47.24, 50.45, 53.58, 56.56, 59.42] 2020-02-01 22:45:18.788: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:45:18.789: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 22:45:18.789: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 22:45:18.789: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.25 2020-02-01 22:45:18.891: INFO @evaluate_confidence: Previous accuracy would be: 92.07 2020-02-01 22:45:18.892: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:45:18.901: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.1, 96.19, 96.27, 96.34, 96.38, 96.46, 96.54, 96.61, 96.67, 96.81, 96.92, 96.98, 97.05, 97.18, 97.24, 97.29, 97.36, 97.42, 97.48, 97.54, 97.61, 97.67, 97.75, 97.82, 97.88, 97.95, 98.0, 98.06, 98.13, 98.16, 98.22, 98.27, 98.28, 98.32, 98.38, 98.47, 98.54, 98.65, 98.71, 98.81, 98.89] 2020-02-01 22:45:18.901: INFO @evaluate_confidence: Dropped ratios are: [14.03, 14.43, 14.83, 15.27, 15.64, 16.02, 16.49, 16.92, 17.33, 17.79, 18.17, 18.61, 19.03, 19.5, 19.88, 20.37, 20.82, 21.34, 21.81, 22.21, 22.59, 23.02, 23.53, 24.09, 24.73, 25.25, 25.74, 26.25, 26.86, 27.37, 27.88, 28.43, 29.03, 29.6, 30.29, 30.9, 31.67, 32.27, 33.02, 33.87, 34.75] 2020-02-01 22:45:18.908: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:45:18.909: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.10 2020-02-01 22:45:18.909: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.10 2020-02-01 22:45:18.909: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.10 2020-02-01 22:45:19.033: INFO @evaluate_confidence: Previous accuracy would be: 51.51 2020-02-01 22:45:19.033: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-01 22:45:19.035: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [56.47, 56.94, 57.55, 58.68, 59.88] 2020-02-01 22:45:19.035: INFO @evaluate_confidence: Dropped ratios are: [44.86, 49.5, 53.95, 58.67, 63.05] 2020-02-01 22:45:19.088: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:45:19.768: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:45:19.852: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:45:20.302: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:45:20.376: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:45:21.069: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:45:21.152: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 22:45:21.153: INFO @evaluate_gender-age_model: groups 0 3.220161 1 3.820362 2 5.135604 3 5.496793 4 6.109175 5 5.997472 6 6.082923 7 6.740062 Name: errors, dtype: float64 2020-02-01 22:45:21.154: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:45:21.613: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:45:21.673: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 22:45:21.674: INFO @evaluate_gender-age_model: groups 0 5.785999 1 4.556577 2 5.026047 3 5.464135 4 7.689916 5 5.939616 6 8.669225 7 12.581927 Name: errors, dtype: float64 2020-02-01 22:45:21.849: INFO @decay_lr : LR updated to `4.7147878e-05` 2020-02-01 22:45:21.850: INFO @log_profile : T train: 121.513385 2020-02-01 22:45:21.850: INFO @log_profile : T valid: 5.389338 2020-02-01 22:45:21.850: INFO @log_profile : T read data: 2.920601 2020-02-01 22:45:21.851: INFO @log_profile : T hooks: 10.120620 2020-02-01 22:45:21.851: INFO @main_loop : Epoch 150 done 2020-02-01 22:45:21.851: INFO @main_loop : Training epoch 151 2020-02-01 22:47:32.581: INFO @log_variables: train loss nanmean: 0.671216 2020-02-01 22:47:32.581: INFO @log_variables: train age_loss mean: 4.929035 2020-02-01 22:47:32.581: INFO @log_variables: train gender_loss mean: 0.111040 2020-02-01 22:47:32.581: INFO @log_variables: train age_mae mean: 5.403938 2020-02-01 22:47:32.581: INFO @log_variables: train gender_accuracy mean: 0.956629 2020-02-01 22:47:32.581: INFO @log_variables: train gender_confidence/loss nanmean: 0.050831 2020-02-01 22:47:32.581: INFO @log_variables: train gender_confidence/accuracy mean: 0.857237 2020-02-01 22:47:32.581: INFO @log_variables: train age_confidence/loss mean: 0.071346 2020-02-01 22:47:32.582: INFO @log_variables: train age_confidence/accuracy mean: 0.609946 2020-02-01 22:47:32.582: INFO @log_variables: valid loss nanmean: 0.855027 2020-02-01 22:47:32.582: INFO @log_variables: valid age_loss mean: 5.866421 2020-02-01 22:47:32.582: INFO @log_variables: valid gender_loss mean: 0.215035 2020-02-01 22:47:32.582: INFO @log_variables: valid age_mae mean: 6.346014 2020-02-01 22:47:32.582: INFO @log_variables: valid gender_accuracy mean: 0.922048 2020-02-01 22:47:32.582: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055871 2020-02-01 22:47:32.582: INFO @log_variables: valid gender_confidence/accuracy mean: 0.865979 2020-02-01 22:47:32.582: INFO @log_variables: valid age_confidence/loss mean: 0.070358 2020-02-01 22:47:32.582: INFO @log_variables: valid age_confidence/accuracy mean: 0.554623 2020-02-01 22:47:32.582: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:47:32.589: INFO @metrics_hook: train age_mae: 5.404 +-0.031 (110372) 2020-02-01 22:47:32.596: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110372) 2020-02-01 22:47:35.316: INFO @metrics_hook: valid age_mae: 6.346 +-0.091 (17639) 2020-02-01 22:47:35.317: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 22:47:36.957: INFO @decay_lr : LR updated to `4.6912137e-05` 2020-02-01 22:47:36.958: INFO @log_profile : T train: 121.673534 2020-02-01 22:47:36.958: INFO @log_profile : T valid: 5.448835 2020-02-01 22:47:36.958: INFO @log_profile : T read data: 2.907039 2020-02-01 22:47:36.958: INFO @log_profile : T hooks: 5.001383 2020-02-01 22:47:36.958: INFO @main_loop : Epoch 151 done 2020-02-01 22:47:36.958: INFO @main_loop : Training epoch 152 2020-02-01 22:49:46.651: INFO @log_variables: train loss nanmean: 0.669700 2020-02-01 22:49:46.652: INFO @log_variables: train age_loss mean: 4.940949 2020-02-01 22:49:46.652: INFO @log_variables: train gender_loss mean: 0.109015 2020-02-01 22:49:46.652: INFO @log_variables: train age_mae mean: 5.416447 2020-02-01 22:49:46.652: INFO @log_variables: train gender_accuracy mean: 0.957248 2020-02-01 22:49:46.652: INFO @log_variables: train gender_confidence/loss nanmean: 0.050128 2020-02-01 22:49:46.652: INFO @log_variables: train gender_confidence/accuracy mean: 0.860442 2020-02-01 22:49:46.652: INFO @log_variables: train age_confidence/loss mean: 0.071290 2020-02-01 22:49:46.652: INFO @log_variables: train age_confidence/accuracy mean: 0.611373 2020-02-01 22:49:46.652: INFO @log_variables: valid loss nanmean: 0.841385 2020-02-01 22:49:46.652: INFO @log_variables: valid age_loss mean: 5.746244 2020-02-01 22:49:46.652: INFO @log_variables: valid gender_loss mean: 0.212191 2020-02-01 22:49:46.652: INFO @log_variables: valid age_mae mean: 6.225301 2020-02-01 22:49:46.652: INFO @log_variables: valid gender_accuracy mean: 0.923919 2020-02-01 22:49:46.652: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055694 2020-02-01 22:49:46.652: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875163 2020-02-01 22:49:46.652: INFO @log_variables: valid age_confidence/loss mean: 0.070403 2020-02-01 22:49:46.652: INFO @log_variables: valid age_confidence/accuracy mean: 0.570157 2020-02-01 22:49:46.652: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:49:46.660: INFO @metrics_hook: train age_mae: 5.416 +-0.031 (110592) 2020-02-01 22:49:46.667: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110592) 2020-02-01 22:49:49.374: INFO @metrics_hook: valid age_mae: 6.225 +-0.089 (17639) 2020-02-01 22:49:49.375: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 22:49:50.840: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:49:50.841: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:49:50.841: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:49:50.841: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:49:50.965: INFO @evaluate_confidence: Previous accuracy would be: 95.72 2020-02-01 22:49:50.965: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:49:51.027: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.23, 98.3, 98.35, 98.42, 98.46, 98.52, 98.56, 98.63, 98.69, 98.73, 98.78, 98.82, 98.87, 98.91, 98.95, 98.99, 99.02, 99.05, 99.07, 99.11, 99.13, 99.16, 99.19, 99.22, 99.25, 99.28, 99.31, 99.34, 99.36, 99.39, 99.42, 99.44, 99.46, 99.48, 99.49, 99.51, 99.54, 99.55, 99.57, 99.59, 99.61, 99.63, 99.65, 99.67, 99.68, 99.69, 99.71, 99.73, 99.74] 2020-02-01 22:49:51.027: INFO @evaluate_confidence: Dropped ratios are: [10.31, 10.73, 11.11, 11.51, 11.9, 12.27, 12.65, 13.07, 13.46, 13.83, 14.24, 14.61, 15.0, 15.41, 15.78, 16.2, 16.59, 16.96, 17.32, 17.71, 18.15, 18.55, 18.94, 19.36, 19.77, 20.21, 20.61, 21.01, 21.48, 21.93, 22.39, 22.85, 23.35, 23.83, 24.33, 24.8, 25.31, 25.85, 26.36, 26.87, 27.44, 28.01, 28.61, 29.22, 29.85, 30.49, 31.17, 31.89, 32.66] 2020-02-01 22:49:51.075: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:49:51.075: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:49:51.075: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:49:51.075: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 22:49:51.214: INFO @evaluate_confidence: Previous accuracy would be: 59.27 2020-02-01 22:49:51.214: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:49:51.229: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.32, 67.97, 68.67, 69.39, 70.14, 70.87, 71.72] 2020-02-01 22:49:51.229: INFO @evaluate_confidence: Dropped ratios are: [41.08, 44.19, 47.4, 50.4, 53.35, 56.21, 58.98] 2020-02-01 22:49:51.237: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:49:51.237: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 22:49:51.237: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.24 2020-02-01 22:49:51.237: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.24 2020-02-01 22:49:51.340: INFO @evaluate_confidence: Previous accuracy would be: 92.39 2020-02-01 22:49:51.340: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 22:49:51.349: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.27, 96.4, 96.46, 96.59, 96.64, 96.73, 96.79, 96.85, 96.94, 97.03, 97.15, 97.23, 97.31, 97.4, 97.45, 97.55, 97.58, 97.68, 97.7, 97.72, 97.77, 97.82, 97.87, 97.94, 98.0, 98.07, 98.12, 98.21, 98.26, 98.26, 98.3, 98.34, 98.38, 98.46, 98.48, 98.55, 98.63, 98.66, 98.73, 98.83] 2020-02-01 22:49:51.349: INFO @evaluate_confidence: Dropped ratios are: [13.12, 13.48, 13.83, 14.24, 14.62, 15.0, 15.34, 15.75, 16.16, 16.55, 16.98, 17.33, 17.87, 18.36, 18.77, 19.13, 19.55, 20.07, 20.45, 20.86, 21.25, 21.7, 22.2, 22.65, 23.15, 23.53, 24.03, 24.62, 25.16, 25.68, 26.31, 26.78, 27.3, 28.0, 28.59, 29.29, 29.99, 30.65, 31.44, 32.42] 2020-02-01 22:49:51.356: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:49:51.356: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 22:49:51.357: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 22:49:51.357: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 22:49:51.481: INFO @evaluate_confidence: Previous accuracy would be: 53.18 2020-02-01 22:49:51.481: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 22:49:51.482: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.44, 59.43, 60.08] 2020-02-01 22:49:51.482: INFO @evaluate_confidence: Dropped ratios are: [45.9, 50.8, 55.37] 2020-02-01 22:49:51.537: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:50:01.699: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:50:01.786: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:50:02.244: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:50:02.494: INFO @decay_lr : LR updated to `4.6677575e-05` 2020-02-01 22:50:02.495: INFO @log_profile : T train: 121.700997 2020-02-01 22:50:02.495: INFO @log_profile : T valid: 5.414775 2020-02-01 22:50:02.495: INFO @log_profile : T read data: 1.879809 2020-02-01 22:50:02.495: INFO @log_profile : T hooks: 16.464402 2020-02-01 22:50:02.495: INFO @main_loop : Epoch 152 done 2020-02-01 22:50:02.495: INFO @main_loop : Training epoch 153 2020-02-01 22:52:14.038: INFO @log_variables: train loss nanmean: 0.673628 2020-02-01 22:52:14.038: INFO @log_variables: train age_loss mean: 4.956472 2020-02-01 22:52:14.038: INFO @log_variables: train gender_loss mean: 0.110608 2020-02-01 22:52:14.038: INFO @log_variables: train age_mae mean: 5.430987 2020-02-01 22:52:14.038: INFO @log_variables: train gender_accuracy mean: 0.956954 2020-02-01 22:52:14.038: INFO @log_variables: train gender_confidence/loss nanmean: 0.051100 2020-02-01 22:52:14.038: INFO @log_variables: train gender_confidence/accuracy mean: 0.856511 2020-02-01 22:52:14.038: INFO @log_variables: train age_confidence/loss mean: 0.071387 2020-02-01 22:52:14.038: INFO @log_variables: train age_confidence/accuracy mean: 0.610124 2020-02-01 22:52:14.038: INFO @log_variables: valid loss nanmean: 0.833209 2020-02-01 22:52:14.038: INFO @log_variables: valid age_loss mean: 5.745755 2020-02-01 22:52:14.039: INFO @log_variables: valid gender_loss mean: 0.202437 2020-02-01 22:52:14.039: INFO @log_variables: valid age_mae mean: 6.225709 2020-02-01 22:52:14.039: INFO @log_variables: valid gender_accuracy mean: 0.926130 2020-02-01 22:52:14.039: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056305 2020-02-01 22:52:14.039: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876807 2020-02-01 22:52:14.039: INFO @log_variables: valid age_confidence/loss mean: 0.070529 2020-02-01 22:52:14.039: INFO @log_variables: valid age_confidence/accuracy mean: 0.550485 2020-02-01 22:52:14.039: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:52:14.046: INFO @metrics_hook: train age_mae: 5.431 +-0.031 (110371) 2020-02-01 22:52:14.053: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110371) 2020-02-01 22:52:16.841: INFO @metrics_hook: valid age_mae: 6.226 +-0.089 (17639) 2020-02-01 22:52:16.843: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 22:52:18.518: INFO @decay_lr : LR updated to `4.6444187e-05` 2020-02-01 22:52:18.519: INFO @log_profile : T train: 122.629686 2020-02-01 22:52:18.520: INFO @log_profile : T valid: 5.436992 2020-02-01 22:52:18.520: INFO @log_profile : T read data: 2.780979 2020-02-01 22:52:18.520: INFO @log_profile : T hooks: 5.099809 2020-02-01 22:52:18.520: INFO @main_loop : Epoch 153 done 2020-02-01 22:52:18.520: INFO @main_loop : Training epoch 154 2020-02-01 22:54:28.946: INFO @log_variables: train loss nanmean: 0.673243 2020-02-01 22:54:28.946: INFO @log_variables: train age_loss mean: 4.940184 2020-02-01 22:54:28.946: INFO @log_variables: train gender_loss mean: 0.110587 2020-02-01 22:54:28.946: INFO @log_variables: train age_mae mean: 5.414705 2020-02-01 22:54:28.946: INFO @log_variables: train gender_accuracy mean: 0.957489 2020-02-01 22:54:28.946: INFO @log_variables: train gender_confidence/loss nanmean: 0.052093 2020-02-01 22:54:28.946: INFO @log_variables: train gender_confidence/accuracy mean: 0.856322 2020-02-01 22:54:28.946: INFO @log_variables: train age_confidence/loss mean: 0.071509 2020-02-01 22:54:28.947: INFO @log_variables: train age_confidence/accuracy mean: 0.609656 2020-02-01 22:54:28.947: INFO @log_variables: valid loss nanmean: 0.850255 2020-02-01 22:54:28.947: INFO @log_variables: valid age_loss mean: 5.768193 2020-02-01 22:54:28.947: INFO @log_variables: valid gender_loss mean: 0.220680 2020-02-01 22:54:28.947: INFO @log_variables: valid age_mae mean: 6.248355 2020-02-01 22:54:28.947: INFO @log_variables: valid gender_accuracy mean: 0.919950 2020-02-01 22:54:28.947: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054859 2020-02-01 22:54:28.947: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869494 2020-02-01 22:54:28.947: INFO @log_variables: valid age_confidence/loss mean: 0.070396 2020-02-01 22:54:28.947: INFO @log_variables: valid age_confidence/accuracy mean: 0.560179 2020-02-01 22:54:28.947: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:54:28.955: INFO @metrics_hook: train age_mae: 5.415 +-0.031 (110372) 2020-02-01 22:54:28.962: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110372) 2020-02-01 22:54:31.700: INFO @metrics_hook: valid age_mae: 6.248 +-0.089 (17639) 2020-02-01 22:54:31.701: INFO @metrics_hook: valid gender_accuracy: 0.920 +-0.004 (17639) 2020-02-01 22:54:33.179: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:54:33.179: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:54:33.179: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.22 2020-02-01 22:54:33.179: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:54:33.307: INFO @evaluate_confidence: Previous accuracy would be: 95.75 2020-02-01 22:54:33.307: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:54:33.370: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.19, 98.24, 98.31, 98.38, 98.43, 98.46, 98.52, 98.57, 98.62, 98.66, 98.71, 98.77, 98.81, 98.85, 98.89, 98.92, 98.95, 98.99, 99.02, 99.04, 99.08, 99.1, 99.13, 99.15, 99.19, 99.22, 99.25, 99.28, 99.3, 99.32, 99.35, 99.38, 99.41, 99.43, 99.46, 99.48, 99.49, 99.51, 99.52, 99.55, 99.56, 99.58, 99.6, 99.61, 99.62, 99.64, 99.65, 99.66, 99.67] 2020-02-01 22:54:33.370: INFO @evaluate_confidence: Dropped ratios are: [10.51, 10.91, 11.3, 11.71, 12.12, 12.52, 12.92, 13.32, 13.72, 14.11, 14.54, 14.93, 15.34, 15.74, 16.12, 16.49, 16.88, 17.28, 17.69, 18.1, 18.51, 18.89, 19.3, 19.69, 20.1, 20.54, 20.99, 21.4, 21.85, 22.3, 22.75, 23.21, 23.69, 24.14, 24.61, 25.13, 25.67, 26.2, 26.76, 27.34, 27.9, 28.47, 29.11, 29.75, 30.42, 31.08, 31.75, 32.48, 33.29] 2020-02-01 22:54:33.419: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:54:33.419: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:54:33.419: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:54:33.419: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 22:54:33.555: INFO @evaluate_confidence: Previous accuracy would be: 59.52 2020-02-01 22:54:33.555: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:54:33.570: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.31, 67.99, 68.74, 69.47, 70.12, 70.91, 71.69] 2020-02-01 22:54:33.570: INFO @evaluate_confidence: Dropped ratios are: [40.82, 43.92, 47.04, 50.22, 53.22, 56.19, 58.97] 2020-02-01 22:54:33.578: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:54:33.578: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 22:54:33.578: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 22:54:33.578: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:54:33.681: INFO @evaluate_confidence: Previous accuracy would be: 92.00 2020-02-01 22:54:33.681: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 22:54:33.690: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.13, 96.26, 96.33, 96.4, 96.48, 96.53, 96.63, 96.68, 96.73, 96.79, 96.83, 96.94, 97.02, 97.11, 97.2, 97.27, 97.32, 97.36, 97.47, 97.53, 97.56, 97.64, 97.69, 97.75, 97.83, 97.88, 97.94, 98.0, 98.02, 98.08, 98.13, 98.19, 98.23, 98.27, 98.38, 98.45, 98.52, 98.59, 98.7, 98.78, 98.85, 98.88] 2020-02-01 22:54:33.690: INFO @evaluate_confidence: Dropped ratios are: [13.27, 13.69, 14.11, 14.45, 14.81, 15.16, 15.53, 15.88, 16.28, 16.7, 17.1, 17.5, 17.9, 18.32, 18.74, 19.13, 19.62, 20.0, 20.5, 20.91, 21.26, 21.72, 22.13, 22.56, 23.1, 23.53, 23.93, 24.42, 24.98, 25.46, 25.92, 26.47, 27.14, 27.72, 28.31, 29.0, 29.68, 30.4, 31.21, 31.96, 32.79, 33.83] 2020-02-01 22:54:33.698: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:54:33.698: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 22:54:33.698: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 22:54:33.698: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 22:54:33.824: INFO @evaluate_confidence: Previous accuracy would be: 52.91 2020-02-01 22:54:33.824: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 22:54:33.825: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.15, 58.45, 59.11] 2020-02-01 22:54:33.825: INFO @evaluate_confidence: Dropped ratios are: [45.28, 50.18, 55.01] 2020-02-01 22:54:33.880: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:54:34.567: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:54:34.649: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:54:35.100: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:54:35.348: INFO @decay_lr : LR updated to `4.6211968e-05` 2020-02-01 22:54:35.349: INFO @log_profile : T train: 121.406135 2020-02-01 22:54:35.349: INFO @log_profile : T valid: 5.448201 2020-02-01 22:54:35.350: INFO @log_profile : T read data: 2.886770 2020-02-01 22:54:35.350: INFO @log_profile : T hooks: 7.012381 2020-02-01 22:54:35.350: INFO @main_loop : Epoch 154 done 2020-02-01 22:54:35.350: INFO @main_loop : Training epoch 155 2020-02-01 22:56:45.406: INFO @log_variables: train loss nanmean: 0.666621 2020-02-01 22:56:45.406: INFO @log_variables: train age_loss mean: 4.899869 2020-02-01 22:56:45.406: INFO @log_variables: train gender_loss mean: 0.108937 2020-02-01 22:56:45.406: INFO @log_variables: train age_mae mean: 5.374205 2020-02-01 22:56:45.406: INFO @log_variables: train gender_accuracy mean: 0.957348 2020-02-01 22:56:45.406: INFO @log_variables: train gender_confidence/loss nanmean: 0.050569 2020-02-01 22:56:45.406: INFO @log_variables: train gender_confidence/accuracy mean: 0.860306 2020-02-01 22:56:45.406: INFO @log_variables: train age_confidence/loss mean: 0.071576 2020-02-01 22:56:45.406: INFO @log_variables: train age_confidence/accuracy mean: 0.611292 2020-02-01 22:56:45.406: INFO @log_variables: valid loss nanmean: 0.829554 2020-02-01 22:56:45.406: INFO @log_variables: valid age_loss mean: 5.724465 2020-02-01 22:56:45.406: INFO @log_variables: valid gender_loss mean: 0.204207 2020-02-01 22:56:45.406: INFO @log_variables: valid age_mae mean: 6.204116 2020-02-01 22:56:45.406: INFO @log_variables: valid gender_accuracy mean: 0.925336 2020-02-01 22:56:45.406: INFO @log_variables: valid gender_confidence/loss nanmean: 0.052755 2020-02-01 22:56:45.406: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870514 2020-02-01 22:56:45.407: INFO @log_variables: valid age_confidence/loss mean: 0.070750 2020-02-01 22:56:45.407: INFO @log_variables: valid age_confidence/accuracy mean: 0.551562 2020-02-01 22:56:45.407: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:56:45.414: INFO @metrics_hook: train age_mae: 5.374 +-0.031 (110592) 2020-02-01 22:56:45.421: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110592) 2020-02-01 22:56:48.183: INFO @metrics_hook: valid age_mae: 6.204 +-0.089 (17639) 2020-02-01 22:56:48.184: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 22:56:49.809: INFO @decay_lr : LR updated to `4.598091e-05` 2020-02-01 22:56:49.810: INFO @log_profile : T train: 121.936846 2020-02-01 22:56:49.811: INFO @log_profile : T valid: 5.522602 2020-02-01 22:56:49.811: INFO @log_profile : T read data: 1.894333 2020-02-01 22:56:49.811: INFO @log_profile : T hooks: 5.030075 2020-02-01 22:56:49.811: INFO @main_loop : Epoch 155 done 2020-02-01 22:56:49.811: INFO @main_loop : Training epoch 156 2020-02-01 22:59:00.277: INFO @log_variables: train loss nanmean: 0.665919 2020-02-01 22:59:00.278: INFO @log_variables: train age_loss mean: 4.895629 2020-02-01 22:59:00.278: INFO @log_variables: train gender_loss mean: 0.108538 2020-02-01 22:59:00.278: INFO @log_variables: train age_mae mean: 5.370042 2020-02-01 22:59:00.278: INFO @log_variables: train gender_accuracy mean: 0.957544 2020-02-01 22:59:00.278: INFO @log_variables: train gender_confidence/loss nanmean: 0.050759 2020-02-01 22:59:00.278: INFO @log_variables: train gender_confidence/accuracy mean: 0.860236 2020-02-01 22:59:00.278: INFO @log_variables: train age_confidence/loss mean: 0.071433 2020-02-01 22:59:00.278: INFO @log_variables: train age_confidence/accuracy mean: 0.610544 2020-02-01 22:59:00.278: INFO @log_variables: valid loss nanmean: 0.839238 2020-02-01 22:59:00.278: INFO @log_variables: valid age_loss mean: 5.722360 2020-02-01 22:59:00.278: INFO @log_variables: valid gender_loss mean: 0.212428 2020-02-01 22:59:00.278: INFO @log_variables: valid age_mae mean: 6.201741 2020-02-01 22:59:00.278: INFO @log_variables: valid gender_accuracy mean: 0.922048 2020-02-01 22:59:00.278: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055214 2020-02-01 22:59:00.278: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867396 2020-02-01 22:59:00.278: INFO @log_variables: valid age_confidence/loss mean: 0.070693 2020-02-01 22:59:00.279: INFO @log_variables: valid age_confidence/accuracy mean: 0.557968 2020-02-01 22:59:00.279: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 22:59:00.286: INFO @metrics_hook: train age_mae: 5.370 +-0.031 (110372) 2020-02-01 22:59:00.293: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 22:59:03.028: INFO @metrics_hook: valid age_mae: 6.202 +-0.088 (17639) 2020-02-01 22:59:03.029: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 22:59:04.494: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:59:04.495: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 22:59:04.495: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 22:59:04.495: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 22:59:04.623: INFO @evaluate_confidence: Previous accuracy would be: 95.75 2020-02-01 22:59:04.624: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 22:59:04.686: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.23, 98.3, 98.35, 98.42, 98.47, 98.52, 98.56, 98.62, 98.68, 98.73, 98.76, 98.8, 98.85, 98.89, 98.93, 98.96, 98.99, 99.04, 99.07, 99.11, 99.14, 99.16, 99.19, 99.21, 99.23, 99.26, 99.3, 99.33, 99.36, 99.38, 99.4, 99.43, 99.45, 99.47, 99.49, 99.51, 99.53, 99.54, 99.56, 99.59, 99.61, 99.63, 99.65, 99.67, 99.68, 99.69, 99.7, 99.72, 99.74] 2020-02-01 22:59:04.686: INFO @evaluate_confidence: Dropped ratios are: [10.38, 10.77, 11.16, 11.56, 11.93, 12.31, 12.68, 13.07, 13.46, 13.88, 14.22, 14.62, 15.03, 15.43, 15.8, 16.17, 16.54, 16.93, 17.34, 17.74, 18.13, 18.53, 18.93, 19.35, 19.77, 20.16, 20.58, 21.03, 21.5, 21.93, 22.31, 22.77, 23.28, 23.79, 24.31, 24.81, 25.33, 25.89, 26.42, 26.96, 27.53, 28.15, 28.77, 29.39, 30.05, 30.68, 31.37, 32.11, 32.9] 2020-02-01 22:59:04.734: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:59:04.735: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.57 +- 0.14 2020-02-01 22:59:04.735: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 22:59:04.735: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 22:59:04.870: INFO @evaluate_confidence: Previous accuracy would be: 59.55 2020-02-01 22:59:04.871: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56] 2020-02-01 22:59:04.885: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.26, 68.01, 68.74, 69.43, 70.21, 70.89, 71.63] 2020-02-01 22:59:04.886: INFO @evaluate_confidence: Dropped ratios are: [40.31, 43.56, 46.71, 49.86, 53.01, 55.91, 58.72] 2020-02-01 22:59:04.893: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 22:59:04.893: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 22:59:04.893: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.25 2020-02-01 22:59:04.894: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 22:59:04.997: INFO @evaluate_confidence: Previous accuracy would be: 92.20 2020-02-01 22:59:04.997: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 22:59:05.006: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.31, 96.42, 96.5, 96.63, 96.72, 96.8, 96.84, 96.9, 97.0, 97.13, 97.16, 97.26, 97.31, 97.39, 97.41, 97.45, 97.52, 97.59, 97.67, 97.71, 97.78, 97.82, 97.88, 97.92, 97.99, 98.02, 98.08, 98.13, 98.17, 98.18, 98.23, 98.27, 98.29, 98.32, 98.4, 98.43, 98.46, 98.5, 98.54, 98.57, 98.7, 98.77] 2020-02-01 22:59:05.006: INFO @evaluate_confidence: Dropped ratios are: [13.39, 13.77, 14.12, 14.52, 14.97, 15.38, 15.74, 16.11, 16.55, 16.91, 17.3, 17.8, 18.26, 18.69, 19.04, 19.45, 19.9, 20.34, 20.8, 21.26, 21.76, 22.14, 22.53, 22.98, 23.44, 23.84, 24.32, 24.76, 25.2, 25.69, 26.17, 26.66, 27.24, 27.85, 28.41, 29.07, 29.73, 30.3, 30.99, 31.74, 32.52, 33.3] 2020-02-01 22:59:05.014: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 22:59:05.014: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 22:59:05.014: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 22:59:05.014: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 22:59:05.142: INFO @evaluate_confidence: Previous accuracy would be: 53.42 2020-02-01 22:59:05.142: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 22:59:05.143: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.59, 58.87, 59.19, 59.79] 2020-02-01 22:59:05.143: INFO @evaluate_confidence: Dropped ratios are: [43.57, 48.06, 52.61, 57.19] 2020-02-01 22:59:05.195: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 22:59:05.880: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:59:05.963: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 22:59:06.426: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 22:59:06.675: INFO @decay_lr : LR updated to `4.5751003e-05` 2020-02-01 22:59:06.677: INFO @log_profile : T train: 121.480206 2020-02-01 22:59:06.677: INFO @log_profile : T valid: 5.450822 2020-02-01 22:59:06.677: INFO @log_profile : T read data: 2.847982 2020-02-01 22:59:06.677: INFO @log_profile : T hooks: 7.008559 2020-02-01 22:59:06.677: INFO @main_loop : Epoch 156 done 2020-02-01 22:59:06.677: INFO @main_loop : Training epoch 157 2020-02-01 23:01:22.715: INFO @log_variables: train loss nanmean: 0.670065 2020-02-01 23:01:22.716: INFO @log_variables: train age_loss mean: 4.928183 2020-02-01 23:01:22.716: INFO @log_variables: train gender_loss mean: 0.109226 2020-02-01 23:01:22.716: INFO @log_variables: train age_mae mean: 5.403021 2020-02-01 23:01:22.716: INFO @log_variables: train gender_accuracy mean: 0.957544 2020-02-01 23:01:22.716: INFO @log_variables: train gender_confidence/loss nanmean: 0.051310 2020-02-01 23:01:22.716: INFO @log_variables: train gender_confidence/accuracy mean: 0.858234 2020-02-01 23:01:22.716: INFO @log_variables: train age_confidence/loss mean: 0.071442 2020-02-01 23:01:22.716: INFO @log_variables: train age_confidence/accuracy mean: 0.606567 2020-02-01 23:01:22.716: INFO @log_variables: valid loss nanmean: 0.826514 2020-02-01 23:01:22.716: INFO @log_variables: valid age_loss mean: 5.684900 2020-02-01 23:01:22.716: INFO @log_variables: valid gender_loss mean: 0.203057 2020-02-01 23:01:22.716: INFO @log_variables: valid age_mae mean: 6.164405 2020-02-01 23:01:22.716: INFO @log_variables: valid gender_accuracy mean: 0.926526 2020-02-01 23:01:22.716: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054333 2020-02-01 23:01:22.716: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873689 2020-02-01 23:01:22.716: INFO @log_variables: valid age_confidence/loss mean: 0.070774 2020-02-01 23:01:22.716: INFO @log_variables: valid age_confidence/accuracy mean: 0.563524 2020-02-01 23:01:22.716: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:01:22.724: INFO @metrics_hook: train age_mae: 5.403 +-0.031 (110372) 2020-02-01 23:01:22.731: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 23:01:25.439: INFO @metrics_hook: valid age_mae: 6.164 +-0.089 (17639) 2020-02-01 23:01:25.440: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 23:01:27.089: INFO @decay_lr : LR updated to `4.5522247e-05` 2020-02-01 23:01:27.090: INFO @log_profile : T train: 126.938799 2020-02-01 23:01:27.090: INFO @log_profile : T valid: 5.483837 2020-02-01 23:01:27.090: INFO @log_profile : T read data: 2.931985 2020-02-01 23:01:27.090: INFO @log_profile : T hooks: 4.983839 2020-02-01 23:01:27.090: INFO @main_loop : Epoch 157 done 2020-02-01 23:01:27.090: INFO @main_loop : Training epoch 158 2020-02-01 23:03:36.973: INFO @log_variables: train loss nanmean: 0.666793 2020-02-01 23:03:36.974: INFO @log_variables: train age_loss mean: 4.897512 2020-02-01 23:03:36.974: INFO @log_variables: train gender_loss mean: 0.108904 2020-02-01 23:03:36.974: INFO @log_variables: train age_mae mean: 5.371930 2020-02-01 23:03:36.974: INFO @log_variables: train gender_accuracy mean: 0.957076 2020-02-01 23:03:36.974: INFO @log_variables: train gender_confidence/loss nanmean: 0.051088 2020-02-01 23:03:36.974: INFO @log_variables: train gender_confidence/accuracy mean: 0.859936 2020-02-01 23:03:36.974: INFO @log_variables: train age_confidence/loss mean: 0.071474 2020-02-01 23:03:36.974: INFO @log_variables: train age_confidence/accuracy mean: 0.609493 2020-02-01 23:03:36.974: INFO @log_variables: valid loss nanmean: 0.838497 2020-02-01 23:03:36.974: INFO @log_variables: valid age_loss mean: 5.771345 2020-02-01 23:03:36.974: INFO @log_variables: valid gender_loss mean: 0.206266 2020-02-01 23:03:36.974: INFO @log_variables: valid age_mae mean: 6.251214 2020-02-01 23:03:36.974: INFO @log_variables: valid gender_accuracy mean: 0.924939 2020-02-01 23:03:36.974: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055820 2020-02-01 23:03:36.974: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871591 2020-02-01 23:03:36.974: INFO @log_variables: valid age_confidence/loss mean: 0.070495 2020-02-01 23:03:36.974: INFO @log_variables: valid age_confidence/accuracy mean: 0.559215 2020-02-01 23:03:36.975: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:03:36.982: INFO @metrics_hook: train age_mae: 5.372 +-0.031 (110592) 2020-02-01 23:03:36.989: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110592) 2020-02-01 23:03:39.735: INFO @metrics_hook: valid age_mae: 6.251 +-0.090 (17639) 2020-02-01 23:03:39.737: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 23:03:41.191: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:03:41.191: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:03:41.191: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 23:03:41.192: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:03:41.319: INFO @evaluate_confidence: Previous accuracy would be: 95.71 2020-02-01 23:03:41.319: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:03:41.380: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.26, 98.33, 98.39, 98.43, 98.5, 98.56, 98.62, 98.67, 98.72, 98.76, 98.79, 98.84, 98.87, 98.91, 98.94, 98.98, 99.02, 99.05, 99.08, 99.11, 99.14, 99.16, 99.18, 99.21, 99.24, 99.27, 99.3, 99.31, 99.34, 99.36, 99.39, 99.4, 99.42, 99.44, 99.45, 99.48, 99.48, 99.5, 99.52, 99.54, 99.57, 99.59, 99.61, 99.62, 99.64, 99.65, 99.67, 99.7, 99.72] 2020-02-01 23:03:41.380: INFO @evaluate_confidence: Dropped ratios are: [10.38, 10.8, 11.21, 11.59, 11.99, 12.39, 12.76, 13.17, 13.55, 13.92, 14.28, 14.71, 15.1, 15.47, 15.84, 16.25, 16.66, 17.06, 17.43, 17.82, 18.22, 18.62, 19.04, 19.45, 19.85, 20.26, 20.65, 21.08, 21.52, 21.97, 22.4, 22.87, 23.31, 23.78, 24.25, 24.76, 25.26, 25.79, 26.38, 26.96, 27.51, 28.09, 28.67, 29.3, 29.94, 30.59, 31.3, 32.04, 32.8] 2020-02-01 23:03:41.429: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:03:41.430: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:03:41.430: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 23:03:41.430: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:03:41.564: INFO @evaluate_confidence: Previous accuracy would be: 59.59 2020-02-01 23:03:41.564: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:03:41.580: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.29, 68.01, 68.66, 69.29, 70.14, 70.94, 71.67, 72.55] 2020-02-01 23:03:41.580: INFO @evaluate_confidence: Dropped ratios are: [40.58, 43.82, 46.93, 49.98, 53.0, 55.94, 58.68, 61.39] 2020-02-01 23:03:41.588: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:03:41.588: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 23:03:41.588: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 23:03:41.588: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:03:41.689: INFO @evaluate_confidence: Previous accuracy would be: 92.49 2020-02-01 23:03:41.689: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:03:41.698: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.38, 96.42, 96.5, 96.58, 96.67, 96.78, 96.83, 96.91, 96.98, 97.04, 97.12, 97.22, 97.26, 97.33, 97.4, 97.51, 97.57, 97.62, 97.65, 97.67, 97.74, 97.8, 97.85, 97.89, 97.97, 98.02, 98.08, 98.14, 98.19, 98.26, 98.28, 98.3, 98.34, 98.4, 98.52, 98.54, 98.58, 98.63, 98.66, 98.74, 98.77, 98.86] 2020-02-01 23:03:41.698: INFO @evaluate_confidence: Dropped ratios are: [12.81, 13.18, 13.6, 14.0, 14.39, 14.86, 15.2, 15.56, 15.96, 16.42, 16.82, 17.23, 17.57, 17.99, 18.33, 18.78, 19.17, 19.58, 20.04, 20.46, 20.89, 21.34, 21.78, 22.23, 22.71, 23.15, 23.69, 24.21, 24.68, 25.1, 25.64, 26.16, 26.69, 27.18, 27.81, 28.49, 29.16, 29.87, 30.61, 31.29, 32.09, 33.13] 2020-02-01 23:03:41.706: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:03:41.706: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 23:03:41.706: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 23:03:41.706: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 23:03:41.831: INFO @evaluate_confidence: Previous accuracy would be: 53.36 2020-02-01 23:03:41.832: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 23:03:41.833: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.87, 59.05, 59.35, 60.24] 2020-02-01 23:03:41.833: INFO @evaluate_confidence: Dropped ratios are: [44.13, 48.71, 53.51, 57.92] 2020-02-01 23:03:41.886: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:03:42.564: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:03:42.645: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:03:43.099: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:03:43.353: INFO @decay_lr : LR updated to `4.5294637e-05` 2020-02-01 23:03:43.355: INFO @log_profile : T train: 121.865668 2020-02-01 23:03:43.355: INFO @log_profile : T valid: 5.438035 2020-02-01 23:03:43.355: INFO @log_profile : T read data: 1.889392 2020-02-01 23:03:43.355: INFO @log_profile : T hooks: 6.995057 2020-02-01 23:03:43.355: INFO @main_loop : Epoch 158 done 2020-02-01 23:03:43.355: INFO @main_loop : Training epoch 159 2020-02-01 23:05:53.960: INFO @log_variables: train loss nanmean: 0.669190 2020-02-01 23:05:53.960: INFO @log_variables: train age_loss mean: 4.914885 2020-02-01 23:05:53.960: INFO @log_variables: train gender_loss mean: 0.109568 2020-02-01 23:05:53.960: INFO @log_variables: train age_mae mean: 5.389794 2020-02-01 23:05:53.960: INFO @log_variables: train gender_accuracy mean: 0.957362 2020-02-01 23:05:53.960: INFO @log_variables: train gender_confidence/loss nanmean: 0.051285 2020-02-01 23:05:53.960: INFO @log_variables: train gender_confidence/accuracy mean: 0.859521 2020-02-01 23:05:53.960: INFO @log_variables: train age_confidence/loss mean: 0.071489 2020-02-01 23:05:53.960: INFO @log_variables: train age_confidence/accuracy mean: 0.612655 2020-02-01 23:05:53.960: INFO @log_variables: valid loss nanmean: 0.844675 2020-02-01 23:05:53.960: INFO @log_variables: valid age_loss mean: 5.699085 2020-02-01 23:05:53.960: INFO @log_variables: valid gender_loss mean: 0.217970 2020-02-01 23:05:53.961: INFO @log_variables: valid age_mae mean: 6.178864 2020-02-01 23:05:53.961: INFO @log_variables: valid gender_accuracy mean: 0.924089 2020-02-01 23:05:53.961: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057764 2020-02-01 23:05:53.961: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875106 2020-02-01 23:05:53.961: INFO @log_variables: valid age_confidence/loss mean: 0.070657 2020-02-01 23:05:53.961: INFO @log_variables: valid age_confidence/accuracy mean: 0.557911 2020-02-01 23:05:53.961: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:05:53.968: INFO @metrics_hook: train age_mae: 5.390 +-0.031 (110372) 2020-02-01 23:05:53.974: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110372) 2020-02-01 23:05:56.703: INFO @metrics_hook: valid age_mae: 6.179 +-0.087 (17639) 2020-02-01 23:05:56.705: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 23:05:58.362: INFO @decay_lr : LR updated to `4.5068165e-05` 2020-02-01 23:05:58.363: INFO @log_profile : T train: 121.656743 2020-02-01 23:05:58.363: INFO @log_profile : T valid: 5.435751 2020-02-01 23:05:58.363: INFO @log_profile : T read data: 2.822627 2020-02-01 23:05:58.364: INFO @log_profile : T hooks: 5.016786 2020-02-01 23:05:58.364: INFO @main_loop : Epoch 159 done 2020-02-01 23:05:58.364: INFO @main_loop : Training epoch 160 2020-02-01 23:08:10.695: INFO @log_variables: train loss nanmean: 0.663545 2020-02-01 23:08:10.695: INFO @log_variables: train age_loss mean: 4.883445 2020-02-01 23:08:10.695: INFO @log_variables: train gender_loss mean: 0.106936 2020-02-01 23:08:10.695: INFO @log_variables: train age_mae mean: 5.358106 2020-02-01 23:08:10.695: INFO @log_variables: train gender_accuracy mean: 0.958531 2020-02-01 23:08:10.695: INFO @log_variables: train gender_confidence/loss nanmean: 0.050793 2020-02-01 23:08:10.695: INFO @log_variables: train gender_confidence/accuracy mean: 0.860553 2020-02-01 23:08:10.695: INFO @log_variables: train age_confidence/loss mean: 0.071588 2020-02-01 23:08:10.695: INFO @log_variables: train age_confidence/accuracy mean: 0.613108 2020-02-01 23:08:10.695: INFO @log_variables: valid loss nanmean: 0.837694 2020-02-01 23:08:10.695: INFO @log_variables: valid age_loss mean: 5.724928 2020-02-01 23:08:10.695: INFO @log_variables: valid gender_loss mean: 0.209520 2020-02-01 23:08:10.695: INFO @log_variables: valid age_mae mean: 6.204295 2020-02-01 23:08:10.695: INFO @log_variables: valid gender_accuracy mean: 0.925563 2020-02-01 23:08:10.696: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056290 2020-02-01 23:08:10.696: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867283 2020-02-01 23:08:10.696: INFO @log_variables: valid age_confidence/loss mean: 0.070483 2020-02-01 23:08:10.696: INFO @log_variables: valid age_confidence/accuracy mean: 0.559385 2020-02-01 23:08:10.696: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:08:10.703: INFO @metrics_hook: train age_mae: 5.358 +-0.031 (110372) 2020-02-01 23:08:10.710: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110372) 2020-02-01 23:08:13.467: INFO @metrics_hook: valid age_mae: 6.204 +-0.089 (17639) 2020-02-01 23:08:13.469: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 23:08:14.961: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:08:14.962: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:08:14.962: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.34 +- 0.21 2020-02-01 23:08:14.962: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:08:15.095: INFO @evaluate_confidence: Previous accuracy would be: 95.85 2020-02-01 23:08:15.096: INFO @evaluate_confidence: Possible optimal thresholds are: [0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:08:15.159: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.32, 98.39, 98.45, 98.49, 98.54, 98.58, 98.64, 98.68, 98.72, 98.76, 98.8, 98.85, 98.89, 98.93, 98.98, 99.02, 99.04, 99.06, 99.1, 99.13, 99.16, 99.19, 99.21, 99.24, 99.26, 99.29, 99.32, 99.35, 99.37, 99.39, 99.42, 99.43, 99.46, 99.47, 99.49, 99.51, 99.53, 99.54, 99.56, 99.58, 99.6, 99.61, 99.62, 99.63, 99.65, 99.66, 99.67, 99.69, 99.71] 2020-02-01 23:08:15.160: INFO @evaluate_confidence: Dropped ratios are: [10.13, 10.51, 10.9, 11.3, 11.69, 12.09, 12.49, 12.87, 13.26, 13.66, 14.04, 14.46, 14.89, 15.26, 15.7, 16.1, 16.49, 16.88, 17.3, 17.72, 18.12, 18.55, 18.97, 19.39, 19.8, 20.23, 20.65, 21.1, 21.56, 22.0, 22.43, 22.91, 23.4, 23.84, 24.33, 24.83, 25.36, 25.87, 26.41, 26.98, 27.56, 28.17, 28.75, 29.34, 30.01, 30.69, 31.39, 32.14, 32.97] 2020-02-01 23:08:15.210: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:08:15.211: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:08:15.211: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 23:08:15.211: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:08:15.353: INFO @evaluate_confidence: Previous accuracy would be: 59.82 2020-02-01 23:08:15.353: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:08:15.370: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.65, 68.19, 68.9, 69.63, 70.36, 71.13, 71.84, 72.59] 2020-02-01 23:08:15.370: INFO @evaluate_confidence: Dropped ratios are: [40.15, 43.33, 46.47, 49.49, 52.5, 55.43, 58.22, 60.95] 2020-02-01 23:08:15.377: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:08:15.378: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 23:08:15.378: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.25 2020-02-01 23:08:15.378: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:08:15.484: INFO @evaluate_confidence: Previous accuracy would be: 92.56 2020-02-01 23:08:15.484: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 23:08:15.493: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.33, 96.4, 96.51, 96.64, 96.72, 96.8, 96.85, 96.93, 96.98, 97.05, 97.1, 97.14, 97.19, 97.28, 97.35, 97.4, 97.49, 97.54, 97.62, 97.68, 97.74, 97.76, 97.79, 97.85, 97.89, 97.95, 97.99, 98.06, 98.1, 98.16, 98.2, 98.25, 98.35, 98.44, 98.52, 98.55, 98.57, 98.64, 98.67, 98.74] 2020-02-01 23:08:15.493: INFO @evaluate_confidence: Dropped ratios are: [13.53, 13.95, 14.34, 14.76, 15.15, 15.54, 15.81, 16.24, 16.59, 16.95, 17.33, 17.76, 18.11, 18.55, 18.91, 19.28, 19.73, 20.17, 20.51, 20.94, 21.35, 21.79, 22.28, 22.8, 23.24, 23.72, 24.2, 24.75, 25.26, 25.8, 26.35, 26.9, 27.65, 28.31, 28.9, 29.51, 30.19, 30.9, 31.58, 32.39] 2020-02-01 23:08:15.500: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:08:15.500: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.10 2020-02-01 23:08:15.501: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:08:15.501: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:08:15.630: INFO @evaluate_confidence: Previous accuracy would be: 53.23 2020-02-01 23:08:15.631: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 23:08:15.632: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.19, 58.56, 59.3, 60.14, 61.28] 2020-02-01 23:08:15.632: INFO @evaluate_confidence: Dropped ratios are: [44.04, 49.05, 53.88, 58.53, 63.03] 2020-02-01 23:08:15.685: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:08:16.421: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:08:16.506: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:08:16.990: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:08:17.067: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:08:17.766: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:08:17.856: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:08:17.858: INFO @evaluate_gender-age_model: groups 0 3.302623 1 3.762148 2 5.004024 3 5.489631 4 6.079167 5 6.020159 6 5.955318 7 6.654976 Name: errors, dtype: float64 2020-02-01 23:08:17.859: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:08:18.319: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:08:18.381: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:08:18.383: INFO @evaluate_gender-age_model: groups 0 5.699848 1 4.929138 2 5.377184 3 5.614971 4 7.380733 5 5.395114 6 7.802142 7 11.674334 Name: errors, dtype: float64 2020-02-01 23:08:18.568: INFO @decay_lr : LR updated to `4.4842825e-05` 2020-02-01 23:08:18.570: INFO @log_profile : T train: 121.371604 2020-02-01 23:08:18.570: INFO @log_profile : T valid: 5.418338 2020-02-01 23:08:18.570: INFO @log_profile : T read data: 2.979372 2020-02-01 23:08:18.570: INFO @log_profile : T hooks: 10.358701 2020-02-01 23:08:18.570: INFO @main_loop : Epoch 160 done 2020-02-01 23:08:18.570: INFO @main_loop : Training epoch 161 2020-02-01 23:10:28.573: INFO @log_variables: train loss nanmean: 0.662963 2020-02-01 23:10:28.573: INFO @log_variables: train age_loss mean: 4.869952 2020-02-01 23:10:28.573: INFO @log_variables: train gender_loss mean: 0.108433 2020-02-01 23:10:28.573: INFO @log_variables: train age_mae mean: 5.344418 2020-02-01 23:10:28.573: INFO @log_variables: train gender_accuracy mean: 0.957357 2020-02-01 23:10:28.573: INFO @log_variables: train gender_confidence/loss nanmean: 0.050189 2020-02-01 23:10:28.573: INFO @log_variables: train gender_confidence/accuracy mean: 0.862250 2020-02-01 23:10:28.573: INFO @log_variables: train age_confidence/loss mean: 0.071475 2020-02-01 23:10:28.574: INFO @log_variables: train age_confidence/accuracy mean: 0.612151 2020-02-01 23:10:28.574: INFO @log_variables: valid loss nanmean: 0.830565 2020-02-01 23:10:28.574: INFO @log_variables: valid age_loss mean: 5.783736 2020-02-01 23:10:28.574: INFO @log_variables: valid gender_loss mean: 0.198077 2020-02-01 23:10:28.574: INFO @log_variables: valid age_mae mean: 6.264627 2020-02-01 23:10:28.574: INFO @log_variables: valid gender_accuracy mean: 0.927944 2020-02-01 23:10:28.574: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054311 2020-02-01 23:10:28.574: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871761 2020-02-01 23:10:28.574: INFO @log_variables: valid age_confidence/loss mean: 0.070390 2020-02-01 23:10:28.574: INFO @log_variables: valid age_confidence/accuracy mean: 0.557798 2020-02-01 23:10:28.574: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:10:28.581: INFO @metrics_hook: train age_mae: 5.344 +-0.031 (110592) 2020-02-01 23:10:28.588: INFO @metrics_hook: train gender_accuracy: 0.957 +-0.001 (110592) 2020-02-01 23:10:31.375: INFO @metrics_hook: valid age_mae: 6.265 +-0.089 (17639) 2020-02-01 23:10:31.377: INFO @metrics_hook: valid gender_accuracy: 0.928 +-0.004 (17639) 2020-02-01 23:10:33.049: INFO @decay_lr : LR updated to `4.461861e-05` 2020-02-01 23:10:33.374: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-01 23:10:33.378: INFO @log_profile : T train: 121.993709 2020-02-01 23:10:33.378: INFO @log_profile : T valid: 5.444167 2020-02-01 23:10:33.378: INFO @log_profile : T read data: 1.881292 2020-02-01 23:10:33.378: INFO @log_profile : T hooks: 5.412347 2020-02-01 23:10:33.378: INFO @main_loop : Epoch 161 done 2020-02-01 23:10:33.378: INFO @main_loop : Training epoch 162 2020-02-01 23:12:44.024: INFO @log_variables: train loss nanmean: 0.664242 2020-02-01 23:12:44.024: INFO @log_variables: train age_loss mean: 4.903717 2020-02-01 23:12:44.024: INFO @log_variables: train gender_loss mean: 0.106794 2020-02-01 23:12:44.024: INFO @log_variables: train age_mae mean: 5.378508 2020-02-01 23:12:44.024: INFO @log_variables: train gender_accuracy mean: 0.958314 2020-02-01 23:12:44.024: INFO @log_variables: train gender_confidence/loss nanmean: 0.049884 2020-02-01 23:12:44.024: INFO @log_variables: train gender_confidence/accuracy mean: 0.860091 2020-02-01 23:12:44.024: INFO @log_variables: train age_confidence/loss mean: 0.071480 2020-02-01 23:12:44.024: INFO @log_variables: train age_confidence/accuracy mean: 0.612211 2020-02-01 23:12:44.024: INFO @log_variables: valid loss nanmean: 0.846543 2020-02-01 23:12:44.024: INFO @log_variables: valid age_loss mean: 5.711551 2020-02-01 23:12:44.024: INFO @log_variables: valid gender_loss mean: 0.219724 2020-02-01 23:12:44.024: INFO @log_variables: valid age_mae mean: 6.190854 2020-02-01 23:12:44.024: INFO @log_variables: valid gender_accuracy mean: 0.922388 2020-02-01 23:12:44.024: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057021 2020-02-01 23:12:44.025: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872839 2020-02-01 23:12:44.025: INFO @log_variables: valid age_confidence/loss mean: 0.070541 2020-02-01 23:12:44.025: INFO @log_variables: valid age_confidence/accuracy mean: 0.555360 2020-02-01 23:12:44.025: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:12:44.032: INFO @metrics_hook: train age_mae: 5.379 +-0.031 (110372) 2020-02-01 23:12:44.039: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 23:12:46.742: INFO @metrics_hook: valid age_mae: 6.191 +-0.087 (17639) 2020-02-01 23:12:46.743: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 23:12:48.196: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:12:48.196: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:12:48.196: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:12:48.197: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:12:48.319: INFO @evaluate_confidence: Previous accuracy would be: 95.83 2020-02-01 23:12:48.319: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:12:48.380: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.24, 98.32, 98.38, 98.45, 98.52, 98.56, 98.6, 98.65, 98.71, 98.75, 98.8, 98.84, 98.88, 98.92, 98.96, 98.99, 99.03, 99.07, 99.09, 99.12, 99.15, 99.19, 99.22, 99.25, 99.28, 99.31, 99.33, 99.36, 99.38, 99.41, 99.43, 99.44, 99.46, 99.48, 99.49, 99.52, 99.54, 99.55, 99.57, 99.59, 99.61, 99.62, 99.63, 99.65, 99.66, 99.68, 99.7, 99.72, 99.73, 99.75] 2020-02-01 23:12:48.381: INFO @evaluate_confidence: Dropped ratios are: [9.86, 10.28, 10.68, 11.09, 11.5, 11.89, 12.28, 12.66, 13.07, 13.45, 13.82, 14.17, 14.54, 14.95, 15.38, 15.81, 16.22, 16.6, 17.01, 17.41, 17.79, 18.17, 18.53, 18.93, 19.32, 19.77, 20.19, 20.6, 21.06, 21.5, 21.94, 22.4, 22.83, 23.3, 23.75, 24.23, 24.71, 25.2, 25.72, 26.26, 26.8, 27.34, 27.94, 28.52, 29.12, 29.77, 30.46, 31.14, 31.87, 32.63] 2020-02-01 23:12:48.430: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:12:48.430: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:12:48.431: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-01 23:12:48.431: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:12:48.564: INFO @evaluate_confidence: Previous accuracy would be: 59.59 2020-02-01 23:12:48.565: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:12:48.581: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.35, 67.98, 68.64, 69.45, 70.16, 70.92, 71.72, 72.63] 2020-02-01 23:12:48.581: INFO @evaluate_confidence: Dropped ratios are: [40.02, 43.17, 46.24, 49.32, 52.44, 55.45, 58.3, 61.01] 2020-02-01 23:12:48.589: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:12:48.589: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 23:12:48.589: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.25 2020-02-01 23:12:48.589: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:12:48.690: INFO @evaluate_confidence: Previous accuracy would be: 92.24 2020-02-01 23:12:48.690: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:12:48.699: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.35, 96.43, 96.49, 96.59, 96.66, 96.7, 96.76, 96.81, 96.93, 97.0, 97.06, 97.12, 97.13, 97.22, 97.29, 97.35, 97.42, 97.45, 97.54, 97.57, 97.63, 97.68, 97.74, 97.77, 97.83, 97.87, 97.93, 98.0, 98.07, 98.14, 98.17, 98.22, 98.26, 98.28, 98.34, 98.4, 98.45, 98.53, 98.57, 98.66] 2020-02-01 23:12:48.699: INFO @evaluate_confidence: Dropped ratios are: [13.41, 13.81, 14.18, 14.65, 15.01, 15.38, 15.78, 16.16, 16.59, 16.88, 17.27, 17.63, 17.99, 18.36, 18.79, 19.22, 19.63, 20.05, 20.42, 20.73, 21.13, 21.58, 21.93, 22.38, 22.95, 23.55, 24.0, 24.48, 25.05, 25.63, 26.15, 26.72, 27.22, 27.83, 28.55, 29.16, 29.87, 30.81, 31.63, 32.54] 2020-02-01 23:12:48.706: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:12:48.706: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.11 2020-02-01 23:12:48.706: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 23:12:48.706: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:12:48.831: INFO @evaluate_confidence: Previous accuracy would be: 53.03 2020-02-01 23:12:48.831: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 23:12:48.833: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.81, 57.95, 58.12, 58.76, 60.13] 2020-02-01 23:12:48.833: INFO @evaluate_confidence: Dropped ratios are: [45.22, 49.73, 54.08, 58.43, 62.66] 2020-02-01 23:12:48.885: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:12:49.583: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:12:49.669: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:12:50.129: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:12:50.385: INFO @decay_lr : LR updated to `4.4395518e-05` 2020-02-01 23:12:50.387: INFO @log_profile : T train: 121.548645 2020-02-01 23:12:50.387: INFO @log_profile : T valid: 5.444349 2020-02-01 23:12:50.387: INFO @log_profile : T read data: 2.963635 2020-02-01 23:12:50.387: INFO @log_profile : T hooks: 6.975889 2020-02-01 23:12:50.387: INFO @main_loop : Epoch 162 done 2020-02-01 23:12:50.387: INFO @main_loop : Training epoch 163 2020-02-01 23:15:01.306: INFO @log_variables: train loss nanmean: 0.662368 2020-02-01 23:15:01.306: INFO @log_variables: train age_loss mean: 4.880624 2020-02-01 23:15:01.306: INFO @log_variables: train gender_loss mean: 0.105776 2020-02-01 23:15:01.306: INFO @log_variables: train age_mae mean: 5.355312 2020-02-01 23:15:01.306: INFO @log_variables: train gender_accuracy mean: 0.958803 2020-02-01 23:15:01.306: INFO @log_variables: train gender_confidence/loss nanmean: 0.050832 2020-02-01 23:15:01.306: INFO @log_variables: train gender_confidence/accuracy mean: 0.861070 2020-02-01 23:15:01.306: INFO @log_variables: train age_confidence/loss mean: 0.071683 2020-02-01 23:15:01.306: INFO @log_variables: train age_confidence/accuracy mean: 0.609067 2020-02-01 23:15:01.306: INFO @log_variables: valid loss nanmean: 0.838525 2020-02-01 23:15:01.306: INFO @log_variables: valid age_loss mean: 5.753335 2020-02-01 23:15:01.307: INFO @log_variables: valid gender_loss mean: 0.209315 2020-02-01 23:15:01.307: INFO @log_variables: valid age_mae mean: 6.233843 2020-02-01 23:15:01.307: INFO @log_variables: valid gender_accuracy mean: 0.923578 2020-02-01 23:15:01.307: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054476 2020-02-01 23:15:01.307: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873462 2020-02-01 23:15:01.307: INFO @log_variables: valid age_confidence/loss mean: 0.070732 2020-02-01 23:15:01.307: INFO @log_variables: valid age_confidence/accuracy mean: 0.559215 2020-02-01 23:15:01.307: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:15:01.314: INFO @metrics_hook: train age_mae: 5.355 +-0.031 (110372) 2020-02-01 23:15:01.321: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110372) 2020-02-01 23:15:04.050: INFO @metrics_hook: valid age_mae: 6.234 +-0.089 (17639) 2020-02-01 23:15:04.051: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 23:15:05.705: INFO @decay_lr : LR updated to `4.417354e-05` 2020-02-01 23:15:05.706: INFO @log_profile : T train: 121.824750 2020-02-01 23:15:05.706: INFO @log_profile : T valid: 5.562768 2020-02-01 23:15:05.706: INFO @log_profile : T read data: 2.867927 2020-02-01 23:15:05.706: INFO @log_profile : T hooks: 4.986057 2020-02-01 23:15:05.706: INFO @main_loop : Epoch 163 done 2020-02-01 23:15:05.706: INFO @main_loop : Training epoch 164 2020-02-01 23:17:15.477: INFO @log_variables: train loss nanmean: 0.663038 2020-02-01 23:17:15.477: INFO @log_variables: train age_loss mean: 4.881203 2020-02-01 23:17:15.477: INFO @log_variables: train gender_loss mean: 0.106421 2020-02-01 23:17:15.477: INFO @log_variables: train age_mae mean: 5.355947 2020-02-01 23:17:15.477: INFO @log_variables: train gender_accuracy mean: 0.958596 2020-02-01 23:17:15.477: INFO @log_variables: train gender_confidence/loss nanmean: 0.050925 2020-02-01 23:17:15.478: INFO @log_variables: train gender_confidence/accuracy mean: 0.861934 2020-02-01 23:17:15.478: INFO @log_variables: train age_confidence/loss mean: 0.071621 2020-02-01 23:17:15.478: INFO @log_variables: train age_confidence/accuracy mean: 0.610966 2020-02-01 23:17:15.478: INFO @log_variables: valid loss nanmean: 0.840126 2020-02-01 23:17:15.478: INFO @log_variables: valid age_loss mean: 5.706335 2020-02-01 23:17:15.478: INFO @log_variables: valid gender_loss mean: 0.216595 2020-02-01 23:17:15.478: INFO @log_variables: valid age_mae mean: 6.185159 2020-02-01 23:17:15.478: INFO @log_variables: valid gender_accuracy mean: 0.920857 2020-02-01 23:17:15.478: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053666 2020-02-01 23:17:15.478: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871988 2020-02-01 23:17:15.478: INFO @log_variables: valid age_confidence/loss mean: 0.070798 2020-02-01 23:17:15.478: INFO @log_variables: valid age_confidence/accuracy mean: 0.556211 2020-02-01 23:17:15.478: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:17:15.485: INFO @metrics_hook: train age_mae: 5.356 +-0.031 (110592) 2020-02-01 23:17:15.493: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110592) 2020-02-01 23:17:18.214: INFO @metrics_hook: valid age_mae: 6.185 +-0.088 (17639) 2020-02-01 23:17:18.215: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-01 23:17:19.673: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:17:19.674: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:17:19.674: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.22 2020-02-01 23:17:19.674: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:17:19.808: INFO @evaluate_confidence: Previous accuracy would be: 95.86 2020-02-01 23:17:19.809: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:17:19.870: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.29, 98.35, 98.42, 98.48, 98.53, 98.57, 98.62, 98.67, 98.71, 98.76, 98.81, 98.84, 98.87, 98.91, 98.94, 98.99, 99.02, 99.07, 99.1, 99.12, 99.14, 99.16, 99.19, 99.22, 99.24, 99.26, 99.28, 99.3, 99.32, 99.34, 99.35, 99.37, 99.39, 99.41, 99.43, 99.45, 99.47, 99.5, 99.51, 99.52, 99.54, 99.56, 99.57, 99.59, 99.62, 99.63, 99.65, 99.67, 99.68, 99.7] 2020-02-01 23:17:19.870: INFO @evaluate_confidence: Dropped ratios are: [9.94, 10.35, 10.75, 11.12, 11.51, 11.9, 12.28, 12.63, 12.98, 13.35, 13.76, 14.12, 14.52, 14.88, 15.25, 15.61, 15.99, 16.39, 16.77, 17.14, 17.53, 17.92, 18.29, 18.69, 19.07, 19.45, 19.87, 20.27, 20.69, 21.1, 21.54, 22.01, 22.45, 22.89, 23.36, 23.84, 24.35, 24.87, 25.4, 25.92, 26.45, 27.0, 27.59, 28.16, 28.74, 29.39, 30.06, 30.71, 31.44, 32.24] 2020-02-01 23:17:19.921: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:17:19.921: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:17:19.921: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:17:19.922: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:17:20.063: INFO @evaluate_confidence: Previous accuracy would be: 59.78 2020-02-01 23:17:20.063: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:17:20.078: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.2, 68.86, 69.6, 70.37, 71.05, 71.85, 72.67] 2020-02-01 23:17:20.078: INFO @evaluate_confidence: Dropped ratios are: [43.29, 46.44, 49.64, 52.7, 55.69, 58.54, 61.14] 2020-02-01 23:17:20.085: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:17:20.086: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.23 2020-02-01 23:17:20.086: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.25 2020-02-01 23:17:20.086: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 23:17:20.192: INFO @evaluate_confidence: Previous accuracy would be: 92.09 2020-02-01 23:17:20.193: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:17:20.202: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.38, 96.5, 96.62, 96.72, 96.8, 96.84, 96.91, 96.99, 97.06, 97.1, 97.16, 97.24, 97.3, 97.35, 97.39, 97.45, 97.48, 97.48, 97.54, 97.6, 97.66, 97.71, 97.78, 97.84, 97.93, 98.01, 98.07, 98.12, 98.15, 98.2, 98.28, 98.31, 98.37, 98.39, 98.43, 98.5, 98.56, 98.61, 98.61, 98.65, 98.7, 98.71, 98.74] 2020-02-01 23:17:20.202: INFO @evaluate_confidence: Dropped ratios are: [13.33, 13.75, 14.15, 14.56, 14.94, 15.26, 15.49, 15.87, 16.2, 16.62, 16.97, 17.39, 17.79, 18.19, 18.54, 18.98, 19.36, 19.71, 20.2, 20.62, 21.03, 21.49, 21.99, 22.44, 22.84, 23.35, 23.7, 24.16, 24.58, 25.03, 25.56, 26.02, 26.62, 27.17, 27.79, 28.39, 28.99, 29.66, 30.33, 30.95, 31.59, 32.39, 33.15] 2020-02-01 23:17:20.210: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:17:20.210: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:17:20.210: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 23:17:20.210: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 23:17:20.338: INFO @evaluate_confidence: Previous accuracy would be: 53.44 2020-02-01 23:17:20.338: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:17:20.339: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.73, 59.26, 60.05] 2020-02-01 23:17:20.340: INFO @evaluate_confidence: Dropped ratios are: [48.05, 52.57, 57.07] 2020-02-01 23:17:20.391: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:17:21.084: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:17:21.168: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:17:21.629: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:17:21.884: INFO @decay_lr : LR updated to `4.395267e-05` 2020-02-01 23:17:21.885: INFO @log_profile : T train: 121.710900 2020-02-01 23:17:21.885: INFO @log_profile : T valid: 5.445124 2020-02-01 23:17:21.885: INFO @log_profile : T read data: 1.913643 2020-02-01 23:17:21.885: INFO @log_profile : T hooks: 7.031061 2020-02-01 23:17:21.885: INFO @main_loop : Epoch 164 done 2020-02-01 23:17:21.885: INFO @main_loop : Training epoch 165 2020-02-01 23:19:32.594: INFO @log_variables: train loss nanmean: 0.661127 2020-02-01 23:19:32.594: INFO @log_variables: train age_loss mean: 4.889730 2020-02-01 23:19:32.594: INFO @log_variables: train gender_loss mean: 0.105345 2020-02-01 23:19:32.594: INFO @log_variables: train age_mae mean: 5.364205 2020-02-01 23:19:32.594: INFO @log_variables: train gender_accuracy mean: 0.958033 2020-02-01 23:19:32.594: INFO @log_variables: train gender_confidence/loss nanmean: 0.049147 2020-02-01 23:19:32.594: INFO @log_variables: train gender_confidence/accuracy mean: 0.864920 2020-02-01 23:19:32.594: INFO @log_variables: train age_confidence/loss mean: 0.071691 2020-02-01 23:19:32.594: INFO @log_variables: train age_confidence/accuracy mean: 0.609511 2020-02-01 23:19:32.595: INFO @log_variables: valid loss nanmean: 0.842020 2020-02-01 23:19:32.595: INFO @log_variables: valid age_loss mean: 5.694060 2020-02-01 23:19:32.595: INFO @log_variables: valid gender_loss mean: 0.217937 2020-02-01 23:19:32.595: INFO @log_variables: valid age_mae mean: 6.173847 2020-02-01 23:19:32.595: INFO @log_variables: valid gender_accuracy mean: 0.922218 2020-02-01 23:19:32.595: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055600 2020-02-01 23:19:32.595: INFO @log_variables: valid gender_confidence/accuracy mean: 0.878961 2020-02-01 23:19:32.595: INFO @log_variables: valid age_confidence/loss mean: 0.070653 2020-02-01 23:19:32.595: INFO @log_variables: valid age_confidence/accuracy mean: 0.560293 2020-02-01 23:19:32.595: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:19:32.602: INFO @metrics_hook: train age_mae: 5.364 +-0.031 (110372) 2020-02-01 23:19:32.608: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 23:19:35.346: INFO @metrics_hook: valid age_mae: 6.174 +-0.088 (17639) 2020-02-01 23:19:35.347: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 23:19:36.969: INFO @decay_lr : LR updated to `4.3732907e-05` 2020-02-01 23:19:36.971: INFO @log_profile : T train: 121.698394 2020-02-01 23:19:36.971: INFO @log_profile : T valid: 5.452506 2020-02-01 23:19:36.971: INFO @log_profile : T read data: 2.856374 2020-02-01 23:19:36.971: INFO @log_profile : T hooks: 5.001025 2020-02-01 23:19:36.971: INFO @main_loop : Epoch 165 done 2020-02-01 23:19:36.971: INFO @main_loop : Training epoch 166 2020-02-01 23:21:47.615: INFO @log_variables: train loss nanmean: 0.659387 2020-02-01 23:21:47.615: INFO @log_variables: train age_loss mean: 4.846089 2020-02-01 23:21:47.615: INFO @log_variables: train gender_loss mean: 0.106507 2020-02-01 23:21:47.615: INFO @log_variables: train age_mae mean: 5.320213 2020-02-01 23:21:47.615: INFO @log_variables: train gender_accuracy mean: 0.958422 2020-02-01 23:21:47.615: INFO @log_variables: train gender_confidence/loss nanmean: 0.050231 2020-02-01 23:21:47.615: INFO @log_variables: train gender_confidence/accuracy mean: 0.861931 2020-02-01 23:21:47.615: INFO @log_variables: train age_confidence/loss mean: 0.071778 2020-02-01 23:21:47.615: INFO @log_variables: train age_confidence/accuracy mean: 0.611305 2020-02-01 23:21:47.615: INFO @log_variables: valid loss nanmean: 0.830060 2020-02-01 23:21:47.615: INFO @log_variables: valid age_loss mean: 5.725573 2020-02-01 23:21:47.615: INFO @log_variables: valid gender_loss mean: 0.200402 2020-02-01 23:21:47.615: INFO @log_variables: valid age_mae mean: 6.205737 2020-02-01 23:21:47.615: INFO @log_variables: valid gender_accuracy mean: 0.927944 2020-02-01 23:21:47.615: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056310 2020-02-01 23:21:47.615: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874880 2020-02-01 23:21:47.616: INFO @log_variables: valid age_confidence/loss mean: 0.071059 2020-02-01 23:21:47.616: INFO @log_variables: valid age_confidence/accuracy mean: 0.557911 2020-02-01 23:21:47.616: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:21:47.623: INFO @metrics_hook: train age_mae: 5.320 +-0.031 (110372) 2020-02-01 23:21:47.630: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 23:21:50.328: INFO @metrics_hook: valid age_mae: 6.206 +-0.089 (17639) 2020-02-01 23:21:50.330: INFO @metrics_hook: valid gender_accuracy: 0.928 +-0.004 (17639) 2020-02-01 23:21:51.781: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:21:51.781: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:21:51.781: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:21:51.781: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:21:51.904: INFO @evaluate_confidence: Previous accuracy would be: 95.84 2020-02-01 23:21:51.904: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:21:51.966: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.31, 98.37, 98.42, 98.48, 98.53, 98.58, 98.62, 98.67, 98.71, 98.76, 98.8, 98.84, 98.88, 98.91, 98.93, 98.98, 99.02, 99.05, 99.08, 99.11, 99.14, 99.17, 99.2, 99.22, 99.24, 99.27, 99.29, 99.31, 99.35, 99.38, 99.41, 99.43, 99.45, 99.47, 99.49, 99.51, 99.52, 99.55, 99.57, 99.59, 99.61, 99.62, 99.63, 99.64, 99.66, 99.68, 99.7, 99.71, 99.72, 99.73] 2020-02-01 23:21:51.966: INFO @evaluate_confidence: Dropped ratios are: [9.8, 10.24, 10.66, 11.07, 11.44, 11.82, 12.17, 12.58, 12.91, 13.29, 13.67, 14.08, 14.46, 14.84, 15.21, 15.6, 15.98, 16.37, 16.75, 17.13, 17.51, 17.91, 18.32, 18.73, 19.13, 19.55, 19.94, 20.35, 20.79, 21.2, 21.64, 22.05, 22.5, 22.97, 23.46, 23.95, 24.45, 24.95, 25.46, 26.01, 26.57, 27.09, 27.69, 28.29, 28.84, 29.47, 30.17, 30.83, 31.59, 32.39] 2020-02-01 23:21:52.016: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:21:52.016: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:21:52.016: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:21:52.016: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:21:52.151: INFO @evaluate_confidence: Previous accuracy would be: 60.12 2020-02-01 23:21:52.151: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:21:52.166: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.39, 69.1, 69.86, 70.52, 71.34, 72.09, 72.85] 2020-02-01 23:21:52.166: INFO @evaluate_confidence: Dropped ratios are: [43.03, 46.17, 49.2, 52.29, 55.3, 58.12, 60.8] 2020-02-01 23:21:52.174: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:21:52.174: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 23:21:52.174: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 23:21:52.174: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:21:52.275: INFO @evaluate_confidence: Previous accuracy would be: 92.79 2020-02-01 23:21:52.275: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:21:52.284: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.58, 96.64, 96.71, 96.8, 96.84, 96.94, 97.04, 97.07, 97.16, 97.18, 97.24, 97.28, 97.36, 97.4, 97.49, 97.54, 97.6, 97.7, 97.75, 97.83, 97.91, 97.95, 97.98, 98.03, 98.11, 98.14, 98.17, 98.22, 98.27, 98.33, 98.37, 98.41, 98.51, 98.54, 98.58, 98.61, 98.64, 98.7, 98.71, 98.76, 98.79, 98.81] 2020-02-01 23:21:52.284: INFO @evaluate_confidence: Dropped ratios are: [12.76, 13.04, 13.4, 13.78, 14.08, 14.49, 14.97, 15.29, 15.74, 16.03, 16.32, 16.75, 17.18, 17.51, 17.9, 18.27, 18.69, 19.16, 19.58, 19.98, 20.39, 20.79, 21.22, 21.64, 22.15, 22.57, 23.09, 23.54, 24.01, 24.55, 25.05, 25.56, 26.24, 26.85, 27.43, 28.04, 28.79, 29.49, 30.38, 31.02, 31.84, 32.71] 2020-02-01 23:21:52.291: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:21:52.291: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:21:52.292: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 23:21:52.292: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:21:52.415: INFO @evaluate_confidence: Previous accuracy would be: 53.72 2020-02-01 23:21:52.415: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:21:52.416: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.84, 59.4, 60.19] 2020-02-01 23:21:52.416: INFO @evaluate_confidence: Dropped ratios are: [46.19, 50.97, 55.8] 2020-02-01 23:21:52.470: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:21:53.155: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:21:53.239: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:21:53.695: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:21:53.958: INFO @decay_lr : LR updated to `4.3514243e-05` 2020-02-01 23:21:53.959: INFO @log_profile : T train: 121.590024 2020-02-01 23:21:53.959: INFO @log_profile : T valid: 5.520726 2020-02-01 23:21:53.960: INFO @log_profile : T read data: 2.834943 2020-02-01 23:21:53.960: INFO @log_profile : T hooks: 6.965159 2020-02-01 23:21:53.960: INFO @main_loop : Epoch 166 done 2020-02-01 23:21:53.960: INFO @main_loop : Training epoch 167 2020-02-01 23:24:04.034: INFO @log_variables: train loss nanmean: 0.658710 2020-02-01 23:24:04.034: INFO @log_variables: train age_loss mean: 4.855280 2020-02-01 23:24:04.034: INFO @log_variables: train gender_loss mean: 0.105013 2020-02-01 23:24:04.034: INFO @log_variables: train age_mae mean: 5.330103 2020-02-01 23:24:04.034: INFO @log_variables: train gender_accuracy mean: 0.958984 2020-02-01 23:24:04.034: INFO @log_variables: train gender_confidence/loss nanmean: 0.050346 2020-02-01 23:24:04.034: INFO @log_variables: train gender_confidence/accuracy mean: 0.861672 2020-02-01 23:24:04.034: INFO @log_variables: train age_confidence/loss mean: 0.071508 2020-02-01 23:24:04.034: INFO @log_variables: train age_confidence/accuracy mean: 0.613100 2020-02-01 23:24:04.034: INFO @log_variables: valid loss nanmean: 0.845170 2020-02-01 23:24:04.035: INFO @log_variables: valid age_loss mean: 5.755324 2020-02-01 23:24:04.035: INFO @log_variables: valid gender_loss mean: 0.213316 2020-02-01 23:24:04.035: INFO @log_variables: valid age_mae mean: 6.235219 2020-02-01 23:24:04.035: INFO @log_variables: valid gender_accuracy mean: 0.924599 2020-02-01 23:24:04.035: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057258 2020-02-01 23:24:04.035: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877261 2020-02-01 23:24:04.035: INFO @log_variables: valid age_confidence/loss mean: 0.070776 2020-02-01 23:24:04.035: INFO @log_variables: valid age_confidence/accuracy mean: 0.557174 2020-02-01 23:24:04.035: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:24:04.042: INFO @metrics_hook: train age_mae: 5.330 +-0.031 (110592) 2020-02-01 23:24:04.049: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110592) 2020-02-01 23:24:06.779: INFO @metrics_hook: valid age_mae: 6.235 +-0.090 (17639) 2020-02-01 23:24:06.780: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 23:24:08.431: INFO @decay_lr : LR updated to `4.329667e-05` 2020-02-01 23:24:08.432: INFO @log_profile : T train: 121.980356 2020-02-01 23:24:08.433: INFO @log_profile : T valid: 5.442861 2020-02-01 23:24:08.433: INFO @log_profile : T read data: 1.954562 2020-02-01 23:24:08.433: INFO @log_profile : T hooks: 5.016674 2020-02-01 23:24:08.433: INFO @main_loop : Epoch 167 done 2020-02-01 23:24:08.433: INFO @main_loop : Training epoch 168 2020-02-01 23:26:19.162: INFO @log_variables: train loss nanmean: 0.656415 2020-02-01 23:26:19.162: INFO @log_variables: train age_loss mean: 4.842466 2020-02-01 23:26:19.162: INFO @log_variables: train gender_loss mean: 0.104095 2020-02-01 23:26:19.162: INFO @log_variables: train age_mae mean: 5.317442 2020-02-01 23:26:19.162: INFO @log_variables: train gender_accuracy mean: 0.959791 2020-02-01 23:26:19.162: INFO @log_variables: train gender_confidence/loss nanmean: 0.049734 2020-02-01 23:26:19.162: INFO @log_variables: train gender_confidence/accuracy mean: 0.862664 2020-02-01 23:26:19.162: INFO @log_variables: train age_confidence/loss mean: 0.071825 2020-02-01 23:26:19.162: INFO @log_variables: train age_confidence/accuracy mean: 0.611994 2020-02-01 23:26:19.163: INFO @log_variables: valid loss nanmean: 0.845095 2020-02-01 23:26:19.163: INFO @log_variables: valid age_loss mean: 5.710540 2020-02-01 23:26:19.163: INFO @log_variables: valid gender_loss mean: 0.218914 2020-02-01 23:26:19.163: INFO @log_variables: valid age_mae mean: 6.190427 2020-02-01 23:26:19.163: INFO @log_variables: valid gender_accuracy mean: 0.922955 2020-02-01 23:26:19.163: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056215 2020-02-01 23:26:19.163: INFO @log_variables: valid gender_confidence/accuracy mean: 0.870798 2020-02-01 23:26:19.163: INFO @log_variables: valid age_confidence/loss mean: 0.070727 2020-02-01 23:26:19.163: INFO @log_variables: valid age_confidence/accuracy mean: 0.559782 2020-02-01 23:26:19.163: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:26:19.170: INFO @metrics_hook: train age_mae: 5.317 +-0.031 (110372) 2020-02-01 23:26:19.177: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110372) 2020-02-01 23:26:21.854: INFO @metrics_hook: valid age_mae: 6.190 +-0.088 (17639) 2020-02-01 23:26:21.855: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 23:26:23.284: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:26:23.284: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:26:23.284: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:26:23.284: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:26:23.407: INFO @evaluate_confidence: Previous accuracy would be: 95.98 2020-02-01 23:26:23.408: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:26:23.468: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.33, 98.38, 98.45, 98.5, 98.55, 98.59, 98.64, 98.7, 98.76, 98.8, 98.87, 98.91, 98.94, 98.97, 99.01, 99.04, 99.07, 99.1, 99.14, 99.17, 99.2, 99.25, 99.27, 99.3, 99.33, 99.36, 99.39, 99.42, 99.44, 99.46, 99.47, 99.5, 99.51, 99.54, 99.56, 99.58, 99.59, 99.61, 99.62, 99.64, 99.65, 99.66, 99.68, 99.7, 99.71, 99.72, 99.73, 99.74, 99.75, 99.76] 2020-02-01 23:26:23.469: INFO @evaluate_confidence: Dropped ratios are: [9.79, 10.19, 10.59, 10.95, 11.31, 11.68, 12.06, 12.45, 12.84, 13.2, 13.59, 13.98, 14.35, 14.72, 15.12, 15.5, 15.85, 16.25, 16.64, 17.01, 17.42, 17.82, 18.2, 18.57, 19.01, 19.39, 19.78, 20.17, 20.6, 21.0, 21.42, 21.84, 22.27, 22.72, 23.19, 23.68, 24.16, 24.67, 25.19, 25.7, 26.21, 26.77, 27.31, 27.93, 28.53, 29.1, 29.74, 30.38, 31.07, 31.83] 2020-02-01 23:26:23.519: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:26:23.519: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:26:23.519: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:26:23.519: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:26:23.652: INFO @evaluate_confidence: Previous accuracy would be: 60.08 2020-02-01 23:26:23.652: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:26:23.667: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.15, 68.81, 69.49, 70.24, 71.01, 71.76, 72.63] 2020-02-01 23:26:23.667: INFO @evaluate_confidence: Dropped ratios are: [42.59, 45.79, 48.97, 51.97, 54.91, 57.82, 60.58] 2020-02-01 23:26:23.674: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:26:23.674: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 23:26:23.674: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 23:26:23.675: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:26:23.777: INFO @evaluate_confidence: Previous accuracy would be: 92.30 2020-02-01 23:26:23.777: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:26:23.786: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.26, 96.35, 96.42, 96.52, 96.63, 96.72, 96.83, 96.92, 96.98, 97.06, 97.15, 97.21, 97.24, 97.32, 97.36, 97.41, 97.45, 97.52, 97.6, 97.64, 97.72, 97.74, 97.8, 97.87, 97.93, 98.0, 98.03, 98.07, 98.1, 98.14, 98.2, 98.25, 98.32, 98.36, 98.44, 98.5, 98.56, 98.6, 98.61, 98.65, 98.72, 98.79] 2020-02-01 23:26:23.786: INFO @evaluate_confidence: Dropped ratios are: [13.12, 13.49, 13.87, 14.29, 14.68, 15.05, 15.47, 15.92, 16.23, 16.63, 16.99, 17.37, 17.74, 18.16, 18.56, 18.87, 19.21, 19.64, 19.98, 20.36, 20.77, 21.21, 21.53, 21.99, 22.36, 22.78, 23.14, 23.62, 24.08, 24.6, 25.15, 25.68, 26.27, 26.78, 27.35, 27.95, 28.59, 29.13, 29.8, 30.65, 31.36, 32.09] 2020-02-01 23:26:23.793: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:26:23.793: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.10 2020-02-01 23:26:23.794: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 23:26:23.794: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:26:23.920: INFO @evaluate_confidence: Previous accuracy would be: 53.17 2020-02-01 23:26:23.920: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-01 23:26:23.921: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.15, 58.33, 58.59, 59.53, 60.1] 2020-02-01 23:26:23.921: INFO @evaluate_confidence: Dropped ratios are: [43.91, 48.81, 53.84, 58.87, 63.25] 2020-02-01 23:26:23.974: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:26:24.666: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:26:24.750: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:26:25.199: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:26:25.460: INFO @decay_lr : LR updated to `4.308019e-05` 2020-02-01 23:26:25.461: INFO @log_profile : T train: 121.488916 2020-02-01 23:26:25.461: INFO @log_profile : T valid: 5.577783 2020-02-01 23:26:25.461: INFO @log_profile : T read data: 2.974226 2020-02-01 23:26:25.461: INFO @log_profile : T hooks: 6.911124 2020-02-01 23:26:25.461: INFO @main_loop : Epoch 168 done 2020-02-01 23:26:25.461: INFO @main_loop : Training epoch 169 2020-02-01 23:28:43.585: INFO @log_variables: train loss nanmean: 0.660851 2020-02-01 23:28:43.585: INFO @log_variables: train age_loss mean: 4.871519 2020-02-01 23:28:43.585: INFO @log_variables: train gender_loss mean: 0.106248 2020-02-01 23:28:43.585: INFO @log_variables: train age_mae mean: 5.346239 2020-02-01 23:28:43.585: INFO @log_variables: train gender_accuracy mean: 0.958395 2020-02-01 23:28:43.585: INFO @log_variables: train gender_confidence/loss nanmean: 0.049760 2020-02-01 23:28:43.586: INFO @log_variables: train gender_confidence/accuracy mean: 0.862103 2020-02-01 23:28:43.586: INFO @log_variables: train age_confidence/loss mean: 0.071637 2020-02-01 23:28:43.586: INFO @log_variables: train age_confidence/accuracy mean: 0.613054 2020-02-01 23:28:43.586: INFO @log_variables: valid loss nanmean: 0.849689 2020-02-01 23:28:43.586: INFO @log_variables: valid age_loss mean: 5.768886 2020-02-01 23:28:43.586: INFO @log_variables: valid gender_loss mean: 0.219212 2020-02-01 23:28:43.586: INFO @log_variables: valid age_mae mean: 6.249923 2020-02-01 23:28:43.586: INFO @log_variables: valid gender_accuracy mean: 0.921708 2020-02-01 23:28:43.586: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055277 2020-02-01 23:28:43.586: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868587 2020-02-01 23:28:43.586: INFO @log_variables: valid age_confidence/loss mean: 0.070685 2020-02-01 23:28:43.586: INFO @log_variables: valid age_confidence/accuracy mean: 0.562277 2020-02-01 23:28:43.586: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:28:43.594: INFO @metrics_hook: train age_mae: 5.346 +-0.031 (110372) 2020-02-01 23:28:43.601: INFO @metrics_hook: train gender_accuracy: 0.958 +-0.001 (110372) 2020-02-01 23:28:46.360: INFO @metrics_hook: valid age_mae: 6.250 +-0.089 (17639) 2020-02-01 23:28:46.361: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 23:28:48.047: INFO @decay_lr : LR updated to `4.2864787e-05` 2020-02-01 23:28:48.048: INFO @log_profile : T train: 128.962469 2020-02-01 23:28:48.048: INFO @log_profile : T valid: 5.605447 2020-02-01 23:28:48.048: INFO @log_profile : T read data: 2.861869 2020-02-01 23:28:48.048: INFO @log_profile : T hooks: 5.081032 2020-02-01 23:28:48.048: INFO @main_loop : Epoch 169 done 2020-02-01 23:28:48.048: INFO @main_loop : Training epoch 170 2020-02-01 23:31:01.114: INFO @log_variables: train loss nanmean: 0.657659 2020-02-01 23:31:01.114: INFO @log_variables: train age_loss mean: 4.830406 2020-02-01 23:31:01.114: INFO @log_variables: train gender_loss mean: 0.105397 2020-02-01 23:31:01.114: INFO @log_variables: train age_mae mean: 5.304736 2020-02-01 23:31:01.114: INFO @log_variables: train gender_accuracy mean: 0.959572 2020-02-01 23:31:01.114: INFO @log_variables: train gender_confidence/loss nanmean: 0.050930 2020-02-01 23:31:01.114: INFO @log_variables: train gender_confidence/accuracy mean: 0.860831 2020-02-01 23:31:01.114: INFO @log_variables: train age_confidence/loss mean: 0.071786 2020-02-01 23:31:01.114: INFO @log_variables: train age_confidence/accuracy mean: 0.611834 2020-02-01 23:31:01.114: INFO @log_variables: valid loss nanmean: 0.826985 2020-02-01 23:31:01.114: INFO @log_variables: valid age_loss mean: 5.702932 2020-02-01 23:31:01.114: INFO @log_variables: valid gender_loss mean: 0.202950 2020-02-01 23:31:01.114: INFO @log_variables: valid age_mae mean: 6.181468 2020-02-01 23:31:01.114: INFO @log_variables: valid gender_accuracy mean: 0.925109 2020-02-01 23:31:01.115: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053402 2020-02-01 23:31:01.115: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871195 2020-02-01 23:31:01.115: INFO @log_variables: valid age_confidence/loss mean: 0.070634 2020-02-01 23:31:01.115: INFO @log_variables: valid age_confidence/accuracy mean: 0.559442 2020-02-01 23:31:01.115: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:31:01.122: INFO @metrics_hook: train age_mae: 5.305 +-0.031 (110592) 2020-02-01 23:31:01.130: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110592) 2020-02-01 23:31:03.860: INFO @metrics_hook: valid age_mae: 6.181 +-0.088 (17639) 2020-02-01 23:31:03.862: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 23:31:05.330: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:31:05.330: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:31:05.330: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:31:05.330: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:31:05.453: INFO @evaluate_confidence: Previous accuracy would be: 95.96 2020-02-01 23:31:05.453: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:31:05.515: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.3, 98.37, 98.42, 98.47, 98.53, 98.58, 98.63, 98.68, 98.72, 98.77, 98.82, 98.86, 98.9, 98.94, 98.98, 99.02, 99.06, 99.08, 99.12, 99.15, 99.19, 99.2, 99.23, 99.25, 99.28, 99.31, 99.34, 99.35, 99.38, 99.4, 99.43, 99.44, 99.46, 99.48, 99.49, 99.51, 99.52, 99.55, 99.57, 99.58, 99.59, 99.59, 99.62, 99.64, 99.66, 99.67, 99.68, 99.69, 99.7, 99.72] 2020-02-01 23:31:05.515: INFO @evaluate_confidence: Dropped ratios are: [9.87, 10.31, 10.68, 11.08, 11.48, 11.89, 12.29, 12.66, 13.02, 13.4, 13.82, 14.19, 14.56, 14.93, 15.31, 15.66, 16.05, 16.42, 16.84, 17.24, 17.66, 18.02, 18.42, 18.83, 19.22, 19.63, 20.02, 20.41, 20.86, 21.31, 21.77, 22.16, 22.61, 23.07, 23.55, 24.0, 24.48, 24.97, 25.47, 25.99, 26.5, 27.03, 27.62, 28.21, 28.82, 29.46, 30.13, 30.82, 31.56, 32.3] 2020-02-01 23:31:05.564: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:31:05.564: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:31:05.564: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:31:05.565: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:31:05.699: INFO @evaluate_confidence: Previous accuracy would be: 60.18 2020-02-01 23:31:05.699: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:31:05.714: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.17, 68.92, 69.7, 70.41, 71.17, 71.94, 72.86] 2020-02-01 23:31:05.714: INFO @evaluate_confidence: Dropped ratios are: [42.1, 45.27, 48.42, 51.44, 54.45, 57.37, 60.18] 2020-02-01 23:31:05.722: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:31:05.722: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 23:31:05.722: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.25 2020-02-01 23:31:05.722: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-01 23:31:05.824: INFO @evaluate_confidence: Previous accuracy would be: 92.51 2020-02-01 23:31:05.825: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 23:31:05.834: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.75, 96.83, 96.9, 96.97, 97.05, 97.12, 97.2, 97.3, 97.34, 97.39, 97.44, 97.5, 97.56, 97.6, 97.64, 97.69, 97.73, 97.78, 97.82, 97.88, 97.91, 97.99, 98.03, 98.06, 98.11, 98.13, 98.15, 98.19, 98.2, 98.23, 98.28, 98.35, 98.39, 98.44, 98.51, 98.57, 98.62, 98.68, 98.69, 98.74, 98.79, 98.87, 98.96] 2020-02-01 23:31:05.834: INFO @evaluate_confidence: Dropped ratios are: [13.4, 13.76, 14.07, 14.4, 14.72, 15.06, 15.45, 15.82, 16.12, 16.46, 16.86, 17.26, 17.63, 17.93, 18.32, 18.68, 19.02, 19.43, 19.8, 20.19, 20.65, 21.15, 21.55, 22.0, 22.46, 22.89, 23.32, 23.72, 24.12, 24.55, 25.05, 25.57, 26.08, 26.57, 27.11, 27.67, 28.24, 28.88, 29.49, 30.15, 30.84, 31.74, 32.66] 2020-02-01 23:31:05.841: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:31:05.842: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:31:05.842: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:31:05.842: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:31:05.965: INFO @evaluate_confidence: Previous accuracy would be: 53.38 2020-02-01 23:31:05.966: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:31:05.967: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.55, 58.72, 58.98] 2020-02-01 23:31:05.967: INFO @evaluate_confidence: Dropped ratios are: [45.43, 49.84, 54.37] 2020-02-01 23:31:06.020: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:31:06.698: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:31:06.782: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:31:07.237: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:31:07.314: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:31:08.017: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:31:08.102: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:31:08.104: INFO @evaluate_gender-age_model: groups 0 3.262999 1 3.722502 2 5.035182 3 5.441059 4 5.996498 5 5.874933 6 5.940570 7 6.566327 Name: errors, dtype: float64 2020-02-01 23:31:08.105: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:31:08.574: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:31:08.636: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:31:08.637: INFO @evaluate_gender-age_model: groups 0 5.887558 1 5.480167 2 5.738812 3 5.495249 4 7.148185 5 5.384128 6 7.487853 7 11.576891 Name: errors, dtype: float64 2020-02-01 23:31:08.827: INFO @decay_lr : LR updated to `4.2650463e-05` 2020-02-01 23:31:08.828: INFO @log_profile : T train: 123.097161 2020-02-01 23:31:08.828: INFO @log_profile : T valid: 5.401576 2020-02-01 23:31:08.828: INFO @log_profile : T read data: 1.902111 2020-02-01 23:31:08.828: INFO @log_profile : T hooks: 10.302488 2020-02-01 23:31:08.828: INFO @main_loop : Epoch 170 done 2020-02-01 23:31:08.828: INFO @main_loop : Training epoch 171 2020-02-01 23:33:28.051: INFO @log_variables: train loss nanmean: 0.654867 2020-02-01 23:33:28.052: INFO @log_variables: train age_loss mean: 4.819003 2020-02-01 23:33:28.052: INFO @log_variables: train gender_loss mean: 0.104981 2020-02-01 23:33:28.052: INFO @log_variables: train age_mae mean: 5.293830 2020-02-01 23:33:28.052: INFO @log_variables: train gender_accuracy mean: 0.958703 2020-02-01 23:33:28.052: INFO @log_variables: train gender_confidence/loss nanmean: 0.049463 2020-02-01 23:33:28.052: INFO @log_variables: train gender_confidence/accuracy mean: 0.862148 2020-02-01 23:33:28.052: INFO @log_variables: train age_confidence/loss mean: 0.071876 2020-02-01 23:33:28.052: INFO @log_variables: train age_confidence/accuracy mean: 0.611967 2020-02-01 23:33:28.052: INFO @log_variables: valid loss nanmean: 0.838835 2020-02-01 23:33:28.052: INFO @log_variables: valid age_loss mean: 5.729933 2020-02-01 23:33:28.052: INFO @log_variables: valid gender_loss mean: 0.211031 2020-02-01 23:33:28.052: INFO @log_variables: valid age_mae mean: 6.208849 2020-02-01 23:33:28.052: INFO @log_variables: valid gender_accuracy mean: 0.923408 2020-02-01 23:33:28.052: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055136 2020-02-01 23:33:28.052: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873292 2020-02-01 23:33:28.052: INFO @log_variables: valid age_confidence/loss mean: 0.070950 2020-02-01 23:33:28.052: INFO @log_variables: valid age_confidence/accuracy mean: 0.560463 2020-02-01 23:33:28.052: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:33:28.059: INFO @metrics_hook: train age_mae: 5.294 +-0.030 (110372) 2020-02-01 23:33:28.066: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110372) 2020-02-01 23:33:30.723: INFO @metrics_hook: valid age_mae: 6.209 +-0.089 (17639) 2020-02-01 23:33:30.724: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 23:33:32.385: INFO @decay_lr : LR updated to `4.2437212e-05` 2020-02-01 23:33:32.386: INFO @log_profile : T train: 130.016014 2020-02-01 23:33:32.386: INFO @log_profile : T valid: 5.657751 2020-02-01 23:33:32.386: INFO @log_profile : T read data: 2.854857 2020-02-01 23:33:32.386: INFO @log_profile : T hooks: 4.953418 2020-02-01 23:33:32.386: INFO @main_loop : Epoch 171 done 2020-02-01 23:33:32.386: INFO @main_loop : Training epoch 172 2020-02-01 23:35:52.919: INFO @log_variables: train loss nanmean: 0.659788 2020-02-01 23:35:52.919: INFO @log_variables: train age_loss mean: 4.861610 2020-02-01 23:35:52.919: INFO @log_variables: train gender_loss mean: 0.105203 2020-02-01 23:35:52.919: INFO @log_variables: train age_mae mean: 5.335955 2020-02-01 23:35:52.919: INFO @log_variables: train gender_accuracy mean: 0.959356 2020-02-01 23:35:52.920: INFO @log_variables: train gender_confidence/loss nanmean: 0.050554 2020-02-01 23:35:52.920: INFO @log_variables: train gender_confidence/accuracy mean: 0.861324 2020-02-01 23:35:52.920: INFO @log_variables: train age_confidence/loss mean: 0.071631 2020-02-01 23:35:52.920: INFO @log_variables: train age_confidence/accuracy mean: 0.614322 2020-02-01 23:35:52.920: INFO @log_variables: valid loss nanmean: 0.848181 2020-02-01 23:35:52.920: INFO @log_variables: valid age_loss mean: 5.725931 2020-02-01 23:35:52.920: INFO @log_variables: valid gender_loss mean: 0.219952 2020-02-01 23:35:52.920: INFO @log_variables: valid age_mae mean: 6.206568 2020-02-01 23:35:52.920: INFO @log_variables: valid gender_accuracy mean: 0.921708 2020-02-01 23:35:52.920: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057324 2020-02-01 23:35:52.920: INFO @log_variables: valid gender_confidence/accuracy mean: 0.868700 2020-02-01 23:35:52.920: INFO @log_variables: valid age_confidence/loss mean: 0.070362 2020-02-01 23:35:52.920: INFO @log_variables: valid age_confidence/accuracy mean: 0.554623 2020-02-01 23:35:52.920: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:35:52.927: INFO @metrics_hook: train age_mae: 5.336 +-0.031 (110372) 2020-02-01 23:35:52.934: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110372) 2020-02-01 23:35:55.656: INFO @metrics_hook: valid age_mae: 6.207 +-0.087 (17639) 2020-02-01 23:35:55.658: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-01 23:35:57.108: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:35:57.108: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:35:57.108: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:35:57.109: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:35:57.233: INFO @evaluate_confidence: Previous accuracy would be: 95.94 2020-02-01 23:35:57.234: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:35:57.296: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.36, 98.41, 98.48, 98.53, 98.58, 98.62, 98.66, 98.7, 98.75, 98.78, 98.81, 98.85, 98.89, 98.92, 98.95, 98.98, 99.02, 99.05, 99.09, 99.13, 99.14, 99.18, 99.22, 99.24, 99.26, 99.29, 99.33, 99.36, 99.38, 99.4, 99.42, 99.44, 99.45, 99.49, 99.51, 99.53, 99.54, 99.56, 99.58, 99.6, 99.61, 99.63, 99.64, 99.66, 99.68, 99.69, 99.7, 99.71, 99.72, 99.74] 2020-02-01 23:35:57.296: INFO @evaluate_confidence: Dropped ratios are: [9.91, 10.28, 10.68, 11.07, 11.47, 11.86, 12.24, 12.63, 13.0, 13.36, 13.73, 14.1, 14.44, 14.81, 15.17, 15.55, 15.97, 16.34, 16.72, 17.07, 17.47, 17.85, 18.25, 18.65, 19.04, 19.43, 19.83, 20.25, 20.67, 21.09, 21.54, 21.98, 22.42, 22.9, 23.31, 23.78, 24.26, 24.77, 25.27, 25.8, 26.34, 26.85, 27.43, 28.04, 28.62, 29.23, 29.87, 30.52, 31.24, 31.97] 2020-02-01 23:35:57.347: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:35:57.347: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:35:57.347: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:35:57.348: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:35:57.486: INFO @evaluate_confidence: Previous accuracy would be: 59.87 2020-02-01 23:35:57.487: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:35:57.502: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.1, 68.72, 69.5, 70.22, 71.04, 71.88, 72.76] 2020-02-01 23:35:57.502: INFO @evaluate_confidence: Dropped ratios are: [42.44, 45.62, 48.76, 51.97, 54.97, 57.79, 60.57] 2020-02-01 23:35:57.509: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:35:57.509: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.22 2020-02-01 23:35:57.510: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.25 2020-02-01 23:35:57.510: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-01 23:35:57.613: INFO @evaluate_confidence: Previous accuracy would be: 92.17 2020-02-01 23:35:57.613: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 23:35:57.622: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.28, 96.34, 96.44, 96.49, 96.62, 96.72, 96.75, 96.82, 96.88, 96.92, 96.97, 97.04, 97.09, 97.18, 97.25, 97.32, 97.37, 97.43, 97.46, 97.51, 97.61, 97.67, 97.75, 97.8, 97.85, 97.91, 97.98, 98.01, 98.08, 98.15, 98.19, 98.23, 98.29, 98.31, 98.32, 98.41, 98.45, 98.51, 98.55, 98.57] 2020-02-01 23:35:57.622: INFO @evaluate_confidence: Dropped ratios are: [13.53, 13.92, 14.38, 14.8, 15.23, 15.58, 16.0, 16.4, 16.76, 17.17, 17.61, 18.03, 18.47, 18.88, 19.3, 19.66, 20.01, 20.44, 20.9, 21.34, 21.83, 22.33, 22.77, 23.2, 23.69, 24.16, 24.66, 25.09, 25.69, 26.18, 26.79, 27.31, 27.92, 28.4, 29.03, 29.68, 30.32, 30.95, 31.66, 32.48] 2020-02-01 23:35:57.630: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:35:57.630: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:35:57.630: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:35:57.630: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:35:57.757: INFO @evaluate_confidence: Previous accuracy would be: 52.96 2020-02-01 23:35:57.758: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:35:57.759: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.91, 58.32, 59.22] 2020-02-01 23:35:57.759: INFO @evaluate_confidence: Dropped ratios are: [46.73, 51.63, 56.62] 2020-02-01 23:35:57.813: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:35:58.497: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:35:58.579: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:35:59.029: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:35:59.294: INFO @decay_lr : LR updated to `4.2225027e-05` 2020-02-01 23:35:59.295: INFO @log_profile : T train: 130.800227 2020-02-01 23:35:59.295: INFO @log_profile : T valid: 6.182036 2020-02-01 23:35:59.295: INFO @log_profile : T read data: 2.856118 2020-02-01 23:35:59.295: INFO @log_profile : T hooks: 6.993417 2020-02-01 23:35:59.295: INFO @main_loop : Epoch 172 done 2020-02-01 23:35:59.295: INFO @main_loop : Training epoch 173 2020-02-01 23:38:11.071: INFO @log_variables: train loss nanmean: 0.650518 2020-02-01 23:38:11.072: INFO @log_variables: train age_loss mean: 4.809112 2020-02-01 23:38:11.072: INFO @log_variables: train gender_loss mean: 0.101442 2020-02-01 23:38:11.072: INFO @log_variables: train age_mae mean: 5.283811 2020-02-01 23:38:11.072: INFO @log_variables: train gender_accuracy mean: 0.960603 2020-02-01 23:38:11.072: INFO @log_variables: train gender_confidence/loss nanmean: 0.049341 2020-02-01 23:38:11.072: INFO @log_variables: train gender_confidence/accuracy mean: 0.863996 2020-02-01 23:38:11.072: INFO @log_variables: train age_confidence/loss mean: 0.071765 2020-02-01 23:38:11.072: INFO @log_variables: train age_confidence/accuracy mean: 0.611898 2020-02-01 23:38:11.072: INFO @log_variables: valid loss nanmean: 0.839565 2020-02-01 23:38:11.072: INFO @log_variables: valid age_loss mean: 5.735390 2020-02-01 23:38:11.072: INFO @log_variables: valid gender_loss mean: 0.211043 2020-02-01 23:38:11.072: INFO @log_variables: valid age_mae mean: 6.215032 2020-02-01 23:38:11.072: INFO @log_variables: valid gender_accuracy mean: 0.925506 2020-02-01 23:38:11.072: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055436 2020-02-01 23:38:11.072: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877374 2020-02-01 23:38:11.072: INFO @log_variables: valid age_confidence/loss mean: 0.070872 2020-02-01 23:38:11.073: INFO @log_variables: valid age_confidence/accuracy mean: 0.552072 2020-02-01 23:38:11.073: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:38:11.080: INFO @metrics_hook: train age_mae: 5.284 +-0.030 (110592) 2020-02-01 23:38:11.087: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110592) 2020-02-01 23:38:13.844: INFO @metrics_hook: valid age_mae: 6.215 +-0.088 (17639) 2020-02-01 23:38:13.846: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 23:38:15.507: INFO @decay_lr : LR updated to `4.20139e-05` 2020-02-01 23:38:15.509: INFO @log_profile : T train: 123.568071 2020-02-01 23:38:15.509: INFO @log_profile : T valid: 5.652875 2020-02-01 23:38:15.509: INFO @log_profile : T read data: 1.856848 2020-02-01 23:38:15.509: INFO @log_profile : T hooks: 5.058979 2020-02-01 23:38:15.509: INFO @main_loop : Epoch 173 done 2020-02-01 23:38:15.509: INFO @main_loop : Training epoch 174 2020-02-01 23:40:34.877: INFO @log_variables: train loss nanmean: 0.656440 2020-02-01 23:40:34.877: INFO @log_variables: train age_loss mean: 4.839779 2020-02-01 23:40:34.877: INFO @log_variables: train gender_loss mean: 0.104437 2020-02-01 23:40:34.877: INFO @log_variables: train age_mae mean: 5.314647 2020-02-01 23:40:34.877: INFO @log_variables: train gender_accuracy mean: 0.959482 2020-02-01 23:40:34.877: INFO @log_variables: train gender_confidence/loss nanmean: 0.049714 2020-02-01 23:40:34.877: INFO @log_variables: train gender_confidence/accuracy mean: 0.862320 2020-02-01 23:40:34.877: INFO @log_variables: train age_confidence/loss mean: 0.071803 2020-02-01 23:40:34.877: INFO @log_variables: train age_confidence/accuracy mean: 0.614241 2020-02-01 23:40:34.877: INFO @log_variables: valid loss nanmean: 0.841986 2020-02-01 23:40:34.877: INFO @log_variables: valid age_loss mean: 5.687941 2020-02-01 23:40:34.877: INFO @log_variables: valid gender_loss mean: 0.215067 2020-02-01 23:40:34.877: INFO @log_variables: valid age_mae mean: 6.168500 2020-02-01 23:40:34.877: INFO @log_variables: valid gender_accuracy mean: 0.924599 2020-02-01 23:40:34.878: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058633 2020-02-01 23:40:34.878: INFO @log_variables: valid gender_confidence/accuracy mean: 0.884120 2020-02-01 23:40:34.878: INFO @log_variables: valid age_confidence/loss mean: 0.070753 2020-02-01 23:40:34.878: INFO @log_variables: valid age_confidence/accuracy mean: 0.552696 2020-02-01 23:40:34.878: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:40:34.885: INFO @metrics_hook: train age_mae: 5.315 +-0.031 (110372) 2020-02-01 23:40:34.892: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110372) 2020-02-01 23:40:37.619: INFO @metrics_hook: valid age_mae: 6.168 +-0.087 (17639) 2020-02-01 23:40:37.621: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 23:40:39.074: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:40:39.074: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:40:39.074: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:40:39.074: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:40:39.200: INFO @evaluate_confidence: Previous accuracy would be: 95.95 2020-02-01 23:40:39.201: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:40:39.263: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.32, 98.39, 98.44, 98.5, 98.55, 98.6, 98.64, 98.68, 98.72, 98.76, 98.8, 98.85, 98.89, 98.94, 98.97, 99.0, 99.04, 99.07, 99.1, 99.14, 99.17, 99.2, 99.23, 99.26, 99.28, 99.32, 99.35, 99.37, 99.39, 99.41, 99.44, 99.47, 99.49, 99.51, 99.52, 99.53, 99.55, 99.56, 99.58, 99.6, 99.61, 99.62, 99.64, 99.65, 99.67, 99.68, 99.7, 99.71, 99.72, 99.74] 2020-02-01 23:40:39.264: INFO @evaluate_confidence: Dropped ratios are: [9.86, 10.29, 10.7, 11.07, 11.42, 11.82, 12.23, 12.58, 12.97, 13.35, 13.7, 14.1, 14.44, 14.84, 15.2, 15.57, 15.92, 16.27, 16.67, 17.06, 17.46, 17.87, 18.26, 18.65, 19.03, 19.39, 19.84, 20.29, 20.69, 21.1, 21.59, 22.03, 22.45, 22.89, 23.32, 23.79, 24.28, 24.75, 25.27, 25.82, 26.36, 26.89, 27.49, 28.09, 28.67, 29.27, 29.9, 30.52, 31.24, 32.04] 2020-02-01 23:40:39.313: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:40:39.313: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:40:39.313: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:40:39.313: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:40:39.453: INFO @evaluate_confidence: Previous accuracy would be: 60.18 2020-02-01 23:40:39.453: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:40:39.468: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.45, 69.05, 69.7, 70.44, 71.21, 71.92, 72.77] 2020-02-01 23:40:39.468: INFO @evaluate_confidence: Dropped ratios are: [42.53, 45.57, 48.75, 51.78, 54.78, 57.53, 60.3] 2020-02-01 23:40:39.476: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:40:39.476: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.88 +- 0.21 2020-02-01 23:40:39.476: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.26 2020-02-01 23:40:39.476: INFO @evaluate_confidence: Average confidence of all samples 0.85 +- 0.24 2020-02-01 23:40:39.580: INFO @evaluate_confidence: Previous accuracy would be: 92.46 2020-02-01 23:40:39.580: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86, 0.87] 2020-02-01 23:40:39.588: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.38, 96.43, 96.5, 96.56, 96.6, 96.64, 96.73, 96.82, 96.89, 96.94, 97.0, 97.06, 97.1, 97.15, 97.2, 97.28, 97.37, 97.42, 97.48, 97.56, 97.62, 97.67, 97.72, 97.8, 97.86, 97.94, 98.0, 98.05, 98.11, 98.19, 98.2, 98.25, 98.32, 98.38, 98.47, 98.52, 98.61, 98.67, 98.72] 2020-02-01 23:40:39.588: INFO @evaluate_confidence: Dropped ratios are: [12.7, 12.91, 13.32, 13.75, 14.07, 14.35, 14.83, 15.22, 15.66, 15.99, 16.3, 16.64, 16.94, 17.38, 17.73, 18.16, 18.63, 19.01, 19.34, 19.66, 20.1, 20.6, 21.0, 21.37, 21.76, 22.24, 22.71, 23.13, 23.7, 24.22, 24.79, 25.34, 25.92, 26.67, 27.21, 27.92, 28.61, 29.37, 30.21] 2020-02-01 23:40:39.596: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:40:39.596: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:40:39.597: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:40:39.597: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:40:39.726: INFO @evaluate_confidence: Previous accuracy would be: 53.34 2020-02-01 23:40:39.726: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:40:39.728: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.95, 58.61, 59.16] 2020-02-01 23:40:39.728: INFO @evaluate_confidence: Dropped ratios are: [45.81, 50.71, 56.02] 2020-02-01 23:40:39.782: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:40:40.464: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:40:40.546: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:40:40.994: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:40:41.258: INFO @decay_lr : LR updated to `4.1803833e-05` 2020-02-01 23:40:41.260: INFO @log_profile : T train: 129.398338 2020-02-01 23:40:41.260: INFO @log_profile : T valid: 6.363688 2020-02-01 23:40:41.260: INFO @log_profile : T read data: 2.925501 2020-02-01 23:40:41.260: INFO @log_profile : T hooks: 6.987371 2020-02-01 23:40:41.260: INFO @main_loop : Epoch 174 done 2020-02-01 23:40:41.260: INFO @main_loop : Training epoch 175 2020-02-01 23:43:00.027: INFO @log_variables: train loss nanmean: 0.654283 2020-02-01 23:43:00.028: INFO @log_variables: train age_loss mean: 4.823163 2020-02-01 23:43:00.028: INFO @log_variables: train gender_loss mean: 0.103532 2020-02-01 23:43:00.028: INFO @log_variables: train age_mae mean: 5.297248 2020-02-01 23:43:00.028: INFO @log_variables: train gender_accuracy mean: 0.959691 2020-02-01 23:43:00.028: INFO @log_variables: train gender_confidence/loss nanmean: 0.049644 2020-02-01 23:43:00.028: INFO @log_variables: train gender_confidence/accuracy mean: 0.864449 2020-02-01 23:43:00.028: INFO @log_variables: train age_confidence/loss mean: 0.072049 2020-02-01 23:43:00.028: INFO @log_variables: train age_confidence/accuracy mean: 0.611369 2020-02-01 23:43:00.028: INFO @log_variables: valid loss nanmean: 0.836107 2020-02-01 23:43:00.028: INFO @log_variables: valid age_loss mean: 5.731027 2020-02-01 23:43:00.028: INFO @log_variables: valid gender_loss mean: 0.209058 2020-02-01 23:43:00.028: INFO @log_variables: valid age_mae mean: 6.211202 2020-02-01 23:43:00.028: INFO @log_variables: valid gender_accuracy mean: 0.923465 2020-02-01 23:43:00.028: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054538 2020-02-01 23:43:00.028: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874653 2020-02-01 23:43:00.028: INFO @log_variables: valid age_confidence/loss mean: 0.070513 2020-02-01 23:43:00.028: INFO @log_variables: valid age_confidence/accuracy mean: 0.553886 2020-02-01 23:43:00.029: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:43:00.036: INFO @metrics_hook: train age_mae: 5.297 +-0.031 (110372) 2020-02-01 23:43:00.043: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110372) 2020-02-01 23:43:02.798: INFO @metrics_hook: valid age_mae: 6.211 +-0.088 (17639) 2020-02-01 23:43:02.799: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 23:43:04.495: INFO @decay_lr : LR updated to `4.1594813e-05` 2020-02-01 23:43:04.496: INFO @log_profile : T train: 129.687458 2020-02-01 23:43:04.496: INFO @log_profile : T valid: 5.546770 2020-02-01 23:43:04.496: INFO @log_profile : T read data: 2.885995 2020-02-01 23:43:04.496: INFO @log_profile : T hooks: 5.041916 2020-02-01 23:43:04.496: INFO @main_loop : Epoch 175 done 2020-02-01 23:43:04.496: INFO @main_loop : Training epoch 176 2020-02-01 23:45:14.409: INFO @log_variables: train loss nanmean: 0.653715 2020-02-01 23:45:14.409: INFO @log_variables: train age_loss mean: 4.820928 2020-02-01 23:45:14.409: INFO @log_variables: train gender_loss mean: 0.103929 2020-02-01 23:45:14.409: INFO @log_variables: train age_mae mean: 5.295510 2020-02-01 23:45:14.410: INFO @log_variables: train gender_accuracy mean: 0.959373 2020-02-01 23:45:14.410: INFO @log_variables: train gender_confidence/loss nanmean: 0.049316 2020-02-01 23:45:14.410: INFO @log_variables: train gender_confidence/accuracy mean: 0.863118 2020-02-01 23:45:14.410: INFO @log_variables: train age_confidence/loss mean: 0.071652 2020-02-01 23:45:14.410: INFO @log_variables: train age_confidence/accuracy mean: 0.615198 2020-02-01 23:45:14.410: INFO @log_variables: valid loss nanmean: 0.838035 2020-02-01 23:45:14.410: INFO @log_variables: valid age_loss mean: 5.712633 2020-02-01 23:45:14.410: INFO @log_variables: valid gender_loss mean: 0.212360 2020-02-01 23:45:14.410: INFO @log_variables: valid age_mae mean: 6.192702 2020-02-01 23:45:14.410: INFO @log_variables: valid gender_accuracy mean: 0.924259 2020-02-01 23:45:14.410: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054666 2020-02-01 23:45:14.410: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867113 2020-02-01 23:45:14.410: INFO @log_variables: valid age_confidence/loss mean: 0.070984 2020-02-01 23:45:14.410: INFO @log_variables: valid age_confidence/accuracy mean: 0.555190 2020-02-01 23:45:14.410: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:45:14.417: INFO @metrics_hook: train age_mae: 5.296 +-0.030 (110592) 2020-02-01 23:45:14.424: INFO @metrics_hook: train gender_accuracy: 0.959 +-0.001 (110592) 2020-02-01 23:45:17.184: INFO @metrics_hook: valid age_mae: 6.193 +-0.088 (17639) 2020-02-01 23:45:17.185: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-01 23:45:18.676: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:45:18.677: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:45:18.677: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:45:18.677: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:45:18.806: INFO @evaluate_confidence: Previous accuracy would be: 95.94 2020-02-01 23:45:18.806: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:45:18.869: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.38, 98.43, 98.48, 98.53, 98.58, 98.63, 98.68, 98.73, 98.77, 98.81, 98.85, 98.9, 98.93, 98.97, 99.0, 99.03, 99.06, 99.1, 99.13, 99.17, 99.19, 99.21, 99.23, 99.25, 99.26, 99.29, 99.32, 99.35, 99.36, 99.38, 99.4, 99.42, 99.44, 99.47, 99.49, 99.51, 99.52, 99.54, 99.56, 99.58, 99.6, 99.62, 99.64, 99.66, 99.68, 99.69, 99.71, 99.72, 99.74, 99.75] 2020-02-01 23:45:18.869: INFO @evaluate_confidence: Dropped ratios are: [9.96, 10.33, 10.72, 11.07, 11.43, 11.78, 12.15, 12.55, 12.91, 13.29, 13.66, 14.04, 14.39, 14.75, 15.14, 15.51, 15.87, 16.24, 16.63, 17.02, 17.41, 17.75, 18.13, 18.51, 18.9, 19.29, 19.74, 20.13, 20.53, 20.93, 21.34, 21.77, 22.18, 22.62, 23.08, 23.52, 23.99, 24.47, 24.95, 25.47, 25.98, 26.49, 27.07, 27.68, 28.26, 28.92, 29.56, 30.25, 30.96, 31.64] 2020-02-01 23:45:18.921: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:45:18.921: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:45:18.921: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:45:18.921: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:45:19.064: INFO @evaluate_confidence: Previous accuracy would be: 60.09 2020-02-01 23:45:19.064: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:45:19.079: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.48, 69.16, 70.0, 70.78, 71.52, 72.22, 72.92] 2020-02-01 23:45:19.080: INFO @evaluate_confidence: Dropped ratios are: [42.47, 45.64, 48.78, 51.9, 54.88, 57.73, 60.45] 2020-02-01 23:45:19.087: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:45:19.087: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-01 23:45:19.087: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.26 2020-02-01 23:45:19.088: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-01 23:45:19.195: INFO @evaluate_confidence: Previous accuracy would be: 92.43 2020-02-01 23:45:19.195: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-01 23:45:19.205: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.52, 96.61, 96.69, 96.73, 96.83, 96.93, 97.0, 97.07, 97.11, 97.18, 97.26, 97.3, 97.37, 97.38, 97.43, 97.5, 97.57, 97.59, 97.66, 97.72, 97.75, 97.8, 97.81, 97.84, 97.89, 97.95, 98.0, 98.04, 98.14, 98.16, 98.2, 98.3, 98.39, 98.44, 98.5, 98.51, 98.55, 98.59, 98.67, 98.7, 98.7, 98.72, 98.75] 2020-02-01 23:45:19.205: INFO @evaluate_confidence: Dropped ratios are: [13.4, 13.76, 14.15, 14.48, 14.8, 15.18, 15.55, 15.93, 16.29, 16.68, 17.09, 17.43, 17.78, 18.12, 18.5, 18.86, 19.29, 19.67, 20.13, 20.59, 20.98, 21.41, 21.82, 22.32, 22.77, 23.16, 23.65, 24.08, 24.66, 25.06, 25.48, 26.11, 26.71, 27.37, 28.01, 28.57, 29.04, 29.7, 30.31, 30.98, 31.66, 32.34, 33.12] 2020-02-01 23:45:19.212: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:45:19.213: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-01 23:45:19.213: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.10 2020-02-01 23:45:19.213: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 23:45:19.344: INFO @evaluate_confidence: Previous accuracy would be: 53.68 2020-02-01 23:45:19.344: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-01 23:45:19.346: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.84, 59.01, 59.41, 60.06] 2020-02-01 23:45:19.346: INFO @evaluate_confidence: Dropped ratios are: [44.58, 49.06, 53.8, 58.58] 2020-02-01 23:45:19.400: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:45:20.100: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:45:20.183: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:45:20.677: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:45:20.949: INFO @decay_lr : LR updated to `4.138684e-05` 2020-02-01 23:45:20.951: INFO @log_profile : T train: 121.831783 2020-02-01 23:45:20.951: INFO @log_profile : T valid: 5.510201 2020-02-01 23:45:20.951: INFO @log_profile : T read data: 1.884876 2020-02-01 23:45:20.951: INFO @log_profile : T hooks: 7.151196 2020-02-01 23:45:20.951: INFO @main_loop : Epoch 176 done 2020-02-01 23:45:20.951: INFO @main_loop : Training epoch 177 2020-02-01 23:47:41.472: INFO @log_variables: train loss nanmean: 0.651057 2020-02-01 23:47:41.472: INFO @log_variables: train age_loss mean: 4.800096 2020-02-01 23:47:41.472: INFO @log_variables: train gender_loss mean: 0.102820 2020-02-01 23:47:41.472: INFO @log_variables: train age_mae mean: 5.274457 2020-02-01 23:47:41.472: INFO @log_variables: train gender_accuracy mean: 0.959754 2020-02-01 23:47:41.472: INFO @log_variables: train gender_confidence/loss nanmean: 0.049380 2020-02-01 23:47:41.473: INFO @log_variables: train gender_confidence/accuracy mean: 0.863580 2020-02-01 23:47:41.473: INFO @log_variables: train age_confidence/loss mean: 0.071832 2020-02-01 23:47:41.473: INFO @log_variables: train age_confidence/accuracy mean: 0.614114 2020-02-01 23:47:41.473: INFO @log_variables: valid loss nanmean: 0.834607 2020-02-01 23:47:41.473: INFO @log_variables: valid age_loss mean: 5.766297 2020-02-01 23:47:41.473: INFO @log_variables: valid gender_loss mean: 0.203077 2020-02-01 23:47:41.473: INFO @log_variables: valid age_mae mean: 6.246200 2020-02-01 23:47:41.473: INFO @log_variables: valid gender_accuracy mean: 0.927377 2020-02-01 23:47:41.473: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054840 2020-02-01 23:47:41.473: INFO @log_variables: valid gender_confidence/accuracy mean: 0.880662 2020-02-01 23:47:41.473: INFO @log_variables: valid age_confidence/loss mean: 0.070943 2020-02-01 23:47:41.473: INFO @log_variables: valid age_confidence/accuracy mean: 0.559782 2020-02-01 23:47:41.473: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:47:41.480: INFO @metrics_hook: train age_mae: 5.274 +-0.030 (110372) 2020-02-01 23:47:41.487: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110372) 2020-02-01 23:47:44.187: INFO @metrics_hook: valid age_mae: 6.246 +-0.089 (17639) 2020-02-01 23:47:44.188: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-01 23:47:45.831: INFO @decay_lr : LR updated to `4.1179905e-05` 2020-02-01 23:47:45.832: INFO @log_profile : T train: 130.170332 2020-02-01 23:47:45.832: INFO @log_profile : T valid: 6.748812 2020-02-01 23:47:45.832: INFO @log_profile : T read data: 2.915969 2020-02-01 23:47:45.832: INFO @log_profile : T hooks: 4.969873 2020-02-01 23:47:45.832: INFO @main_loop : Epoch 177 done 2020-02-01 23:47:45.832: INFO @main_loop : Training epoch 178 2020-02-01 23:50:04.836: INFO @log_variables: train loss nanmean: 0.651312 2020-02-01 23:50:04.837: INFO @log_variables: train age_loss mean: 4.801998 2020-02-01 23:50:04.837: INFO @log_variables: train gender_loss mean: 0.101500 2020-02-01 23:50:04.837: INFO @log_variables: train age_mae mean: 5.276131 2020-02-01 23:50:04.837: INFO @log_variables: train gender_accuracy mean: 0.960977 2020-02-01 23:50:04.837: INFO @log_variables: train gender_confidence/loss nanmean: 0.050348 2020-02-01 23:50:04.837: INFO @log_variables: train gender_confidence/accuracy mean: 0.861070 2020-02-01 23:50:04.837: INFO @log_variables: train age_confidence/loss mean: 0.072146 2020-02-01 23:50:04.837: INFO @log_variables: train age_confidence/accuracy mean: 0.609231 2020-02-01 23:50:04.837: INFO @log_variables: valid loss nanmean: 0.836695 2020-02-01 23:50:04.837: INFO @log_variables: valid age_loss mean: 5.741605 2020-02-01 23:50:04.837: INFO @log_variables: valid gender_loss mean: 0.207826 2020-02-01 23:50:04.837: INFO @log_variables: valid age_mae mean: 6.221205 2020-02-01 23:50:04.837: INFO @log_variables: valid gender_accuracy mean: 0.925336 2020-02-01 23:50:04.837: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055210 2020-02-01 23:50:04.837: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872725 2020-02-01 23:50:04.837: INFO @log_variables: valid age_confidence/loss mean: 0.070588 2020-02-01 23:50:04.837: INFO @log_variables: valid age_confidence/accuracy mean: 0.559272 2020-02-01 23:50:04.837: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:50:04.844: INFO @metrics_hook: train age_mae: 5.276 +-0.030 (110372) 2020-02-01 23:50:04.852: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-01 23:50:07.579: INFO @metrics_hook: valid age_mae: 6.221 +-0.088 (17639) 2020-02-01 23:50:07.581: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-01 23:50:09.033: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:50:09.034: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:50:09.034: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.22 2020-02-01 23:50:09.034: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:50:09.160: INFO @evaluate_confidence: Previous accuracy would be: 96.10 2020-02-01 23:50:09.160: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:50:09.223: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.43, 98.48, 98.52, 98.57, 98.63, 98.68, 98.73, 98.77, 98.81, 98.84, 98.89, 98.92, 98.95, 98.99, 99.01, 99.04, 99.07, 99.1, 99.12, 99.15, 99.17, 99.19, 99.22, 99.25, 99.28, 99.31, 99.33, 99.35, 99.38, 99.4, 99.42, 99.44, 99.47, 99.49, 99.5, 99.52, 99.55, 99.56, 99.58, 99.59, 99.6, 99.62, 99.64, 99.66, 99.66, 99.68, 99.7, 99.71, 99.71, 99.72] 2020-02-01 23:50:09.224: INFO @evaluate_confidence: Dropped ratios are: [9.9, 10.23, 10.61, 10.99, 11.38, 11.74, 12.11, 12.48, 12.84, 13.22, 13.61, 14.01, 14.38, 14.72, 15.1, 15.49, 15.88, 16.29, 16.63, 17.0, 17.37, 17.74, 18.1, 18.49, 18.9, 19.28, 19.67, 20.07, 20.47, 20.89, 21.31, 21.77, 22.24, 22.67, 23.14, 23.63, 24.09, 24.58, 25.07, 25.55, 26.06, 26.59, 27.1, 27.63, 28.23, 28.82, 29.49, 30.17, 30.88, 31.56] 2020-02-01 23:50:09.274: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:50:09.274: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:50:09.274: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:50:09.275: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:50:09.416: INFO @evaluate_confidence: Previous accuracy would be: 60.33 2020-02-01 23:50:09.417: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:50:09.432: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [67.96, 68.59, 69.32, 69.98, 70.84, 71.65, 72.62] 2020-02-01 23:50:09.432: INFO @evaluate_confidence: Dropped ratios are: [42.13, 45.39, 48.66, 51.75, 54.76, 57.66, 60.43] 2020-02-01 23:50:09.440: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:50:09.440: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-01 23:50:09.440: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-01 23:50:09.440: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:50:09.548: INFO @evaluate_confidence: Previous accuracy would be: 92.53 2020-02-01 23:50:09.549: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:50:09.558: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.52, 96.63, 96.67, 96.75, 96.82, 96.87, 96.97, 97.03, 97.08, 97.17, 97.24, 97.29, 97.38, 97.45, 97.48, 97.53, 97.6, 97.66, 97.71, 97.76, 97.8, 97.86, 97.91, 97.93, 98.02, 98.09, 98.16, 98.19, 98.22, 98.26, 98.27, 98.32, 98.37, 98.42, 98.48, 98.55, 98.6, 98.66, 98.68, 98.75, 98.79, 98.87] 2020-02-01 23:50:09.558: INFO @evaluate_confidence: Dropped ratios are: [13.16, 13.56, 13.93, 14.2, 14.54, 14.86, 15.26, 15.58, 15.96, 16.35, 16.67, 17.03, 17.39, 17.82, 18.08, 18.48, 18.84, 19.28, 19.71, 20.14, 20.58, 21.04, 21.43, 21.82, 22.25, 22.73, 23.1, 23.53, 24.03, 24.55, 25.07, 25.63, 26.11, 26.59, 27.2, 27.88, 28.53, 29.09, 29.73, 30.39, 31.2, 31.99] 2020-02-01 23:50:09.565: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:50:09.566: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:50:09.566: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:50:09.566: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:50:09.697: INFO @evaluate_confidence: Previous accuracy would be: 52.91 2020-02-01 23:50:09.698: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:50:09.699: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.15, 58.65, 59.4] 2020-02-01 23:50:09.699: INFO @evaluate_confidence: Dropped ratios are: [45.78, 51.51, 56.95] 2020-02-01 23:50:09.753: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:50:10.453: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:50:10.537: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:50:11.013: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:50:11.285: INFO @decay_lr : LR updated to `4.0974006e-05` 2020-02-01 23:50:11.286: INFO @log_profile : T train: 130.028221 2020-02-01 23:50:11.286: INFO @log_profile : T valid: 5.406092 2020-02-01 23:50:11.287: INFO @log_profile : T read data: 2.884466 2020-02-01 23:50:11.287: INFO @log_profile : T hooks: 7.059537 2020-02-01 23:50:11.287: INFO @main_loop : Epoch 178 done 2020-02-01 23:50:11.287: INFO @main_loop : Training epoch 179 2020-02-01 23:52:21.556: INFO @log_variables: train loss nanmean: 0.650651 2020-02-01 23:52:21.556: INFO @log_variables: train age_loss mean: 4.813976 2020-02-01 23:52:21.556: INFO @log_variables: train gender_loss mean: 0.100592 2020-02-01 23:52:21.556: INFO @log_variables: train age_mae mean: 5.288331 2020-02-01 23:52:21.556: INFO @log_variables: train gender_accuracy mean: 0.961326 2020-02-01 23:52:21.556: INFO @log_variables: train gender_confidence/loss nanmean: 0.049606 2020-02-01 23:52:21.557: INFO @log_variables: train gender_confidence/accuracy mean: 0.864809 2020-02-01 23:52:21.557: INFO @log_variables: train age_confidence/loss mean: 0.071964 2020-02-01 23:52:21.557: INFO @log_variables: train age_confidence/accuracy mean: 0.610442 2020-02-01 23:52:21.557: INFO @log_variables: valid loss nanmean: 0.836147 2020-02-01 23:52:21.557: INFO @log_variables: valid age_loss mean: 5.667223 2020-02-01 23:52:21.557: INFO @log_variables: valid gender_loss mean: 0.211950 2020-02-01 23:52:21.557: INFO @log_variables: valid age_mae mean: 6.146924 2020-02-01 23:52:21.557: INFO @log_variables: valid gender_accuracy mean: 0.926300 2020-02-01 23:52:21.557: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057325 2020-02-01 23:52:21.557: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876637 2020-02-01 23:52:21.557: INFO @log_variables: valid age_confidence/loss mean: 0.070939 2020-02-01 23:52:21.557: INFO @log_variables: valid age_confidence/accuracy mean: 0.554397 2020-02-01 23:52:21.557: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:52:21.564: INFO @metrics_hook: train age_mae: 5.288 +-0.030 (110592) 2020-02-01 23:52:21.571: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110592) 2020-02-01 23:52:24.294: INFO @metrics_hook: valid age_mae: 6.147 +-0.088 (17639) 2020-02-01 23:52:24.295: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 23:52:25.958: INFO @decay_lr : LR updated to `4.0769137e-05` 2020-02-01 23:52:25.959: INFO @log_profile : T train: 122.161113 2020-02-01 23:52:25.959: INFO @log_profile : T valid: 5.502780 2020-02-01 23:52:25.959: INFO @log_profile : T read data: 1.903738 2020-02-01 23:52:25.959: INFO @log_profile : T hooks: 5.028494 2020-02-01 23:52:25.959: INFO @main_loop : Epoch 179 done 2020-02-01 23:52:25.959: INFO @main_loop : Training epoch 180 2020-02-01 23:54:38.205: INFO @log_variables: train loss nanmean: 0.650366 2020-02-01 23:54:38.205: INFO @log_variables: train age_loss mean: 4.799359 2020-02-01 23:54:38.205: INFO @log_variables: train gender_loss mean: 0.101417 2020-02-01 23:54:38.205: INFO @log_variables: train age_mae mean: 5.273755 2020-02-01 23:54:38.205: INFO @log_variables: train gender_accuracy mean: 0.960352 2020-02-01 23:54:38.205: INFO @log_variables: train gender_confidence/loss nanmean: 0.049803 2020-02-01 23:54:38.205: INFO @log_variables: train gender_confidence/accuracy mean: 0.863099 2020-02-01 23:54:38.205: INFO @log_variables: train age_confidence/loss mean: 0.072060 2020-02-01 23:54:38.205: INFO @log_variables: train age_confidence/accuracy mean: 0.609521 2020-02-01 23:54:38.205: INFO @log_variables: valid loss nanmean: 0.843972 2020-02-01 23:54:38.205: INFO @log_variables: valid age_loss mean: 5.711822 2020-02-01 23:54:38.205: INFO @log_variables: valid gender_loss mean: 0.217072 2020-02-01 23:54:38.205: INFO @log_variables: valid age_mae mean: 6.191910 2020-02-01 23:54:38.205: INFO @log_variables: valid gender_accuracy mean: 0.923238 2020-02-01 23:54:38.206: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056738 2020-02-01 23:54:38.206: INFO @log_variables: valid gender_confidence/accuracy mean: 0.876467 2020-02-01 23:54:38.206: INFO @log_variables: valid age_confidence/loss mean: 0.070639 2020-02-01 23:54:38.206: INFO @log_variables: valid age_confidence/accuracy mean: 0.561653 2020-02-01 23:54:38.206: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:54:38.213: INFO @metrics_hook: train age_mae: 5.274 +-0.030 (110372) 2020-02-01 23:54:38.221: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110372) 2020-02-01 23:54:40.952: INFO @metrics_hook: valid age_mae: 6.192 +-0.088 (17639) 2020-02-01 23:54:40.953: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 23:54:42.423: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:54:42.423: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-01 23:54:42.423: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-01 23:54:42.424: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-01 23:54:42.548: INFO @evaluate_confidence: Previous accuracy would be: 96.04 2020-02-01 23:54:42.548: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-01 23:54:42.609: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.36, 98.41, 98.47, 98.53, 98.58, 98.64, 98.71, 98.74, 98.78, 98.82, 98.85, 98.91, 98.94, 98.97, 99.02, 99.04, 99.08, 99.11, 99.14, 99.17, 99.2, 99.23, 99.25, 99.27, 99.29, 99.32, 99.35, 99.38, 99.39, 99.41, 99.44, 99.46, 99.48, 99.49, 99.51, 99.53, 99.54, 99.56, 99.57, 99.59, 99.6, 99.61, 99.63, 99.65, 99.67, 99.68, 99.7, 99.71, 99.73, 99.74] 2020-02-01 23:54:42.610: INFO @evaluate_confidence: Dropped ratios are: [9.71, 10.09, 10.45, 10.88, 11.26, 11.62, 12.0, 12.37, 12.72, 13.09, 13.43, 13.82, 14.2, 14.62, 15.02, 15.39, 15.79, 16.16, 16.57, 16.96, 17.32, 17.7, 18.08, 18.44, 18.82, 19.21, 19.64, 20.03, 20.43, 20.85, 21.28, 21.71, 22.12, 22.56, 23.0, 23.48, 23.98, 24.48, 25.0, 25.53, 26.09, 26.64, 27.23, 27.8, 28.38, 28.99, 29.64, 30.3, 31.02, 31.76] 2020-02-01 23:54:42.659: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:54:42.659: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:54:42.659: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.11 2020-02-01 23:54:42.659: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:54:42.797: INFO @evaluate_confidence: Previous accuracy would be: 60.33 2020-02-01 23:54:42.797: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:54:42.812: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.25, 68.91, 69.58, 70.39, 71.23, 72.07, 72.86] 2020-02-01 23:54:42.812: INFO @evaluate_confidence: Dropped ratios are: [42.45, 45.71, 48.92, 52.12, 55.12, 58.02, 60.74] 2020-02-01 23:54:42.820: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:54:42.820: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.87 +- 0.22 2020-02-01 23:54:42.820: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.25 2020-02-01 23:54:42.821: INFO @evaluate_confidence: Average confidence of all samples 0.84 +- 0.25 2020-02-01 23:54:42.924: INFO @evaluate_confidence: Previous accuracy would be: 92.32 2020-02-01 23:54:42.925: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-01 23:54:42.933: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.27, 96.33, 96.41, 96.53, 96.59, 96.64, 96.73, 96.78, 96.83, 96.92, 96.99, 97.08, 97.14, 97.18, 97.28, 97.34, 97.41, 97.48, 97.56, 97.62, 97.69, 97.73, 97.79, 97.89, 97.92, 98.0, 98.04, 98.12, 98.16, 98.22, 98.24, 98.29, 98.35, 98.4, 98.46, 98.5, 98.6, 98.65, 98.73, 98.76, 98.81] 2020-02-01 23:54:42.933: INFO @evaluate_confidence: Dropped ratios are: [12.71, 13.03, 13.45, 13.86, 14.17, 14.57, 14.99, 15.36, 15.74, 16.19, 16.51, 16.87, 17.26, 17.63, 18.02, 18.33, 18.66, 19.08, 19.46, 19.87, 20.23, 20.59, 21.01, 21.45, 21.92, 22.37, 22.79, 23.22, 23.72, 24.3, 24.82, 25.35, 25.93, 26.42, 27.01, 27.74, 28.37, 28.99, 29.69, 30.43, 31.14] 2020-02-01 23:54:42.941: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:54:42.941: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:54:42.941: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-01 23:54:42.941: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-01 23:54:43.066: INFO @evaluate_confidence: Previous accuracy would be: 53.43 2020-02-01 23:54:43.066: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:54:43.067: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [59.05, 59.63, 60.45] 2020-02-01 23:54:43.068: INFO @evaluate_confidence: Dropped ratios are: [46.95, 51.77, 56.8] 2020-02-01 23:54:43.118: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:54:43.825: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:54:43.908: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:54:44.377: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:54:44.454: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:54:45.157: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:54:45.240: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:54:45.242: INFO @evaluate_gender-age_model: groups 0 3.237461 1 3.740415 2 4.927731 3 5.375213 4 5.991268 5 5.874979 6 5.894624 7 6.481139 Name: errors, dtype: float64 2020-02-01 23:54:45.243: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:54:45.711: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:54:45.774: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-01 23:54:45.775: INFO @evaluate_gender-age_model: groups 0 6.074611 1 5.238400 2 5.567932 3 5.818173 4 7.437743 5 5.407702 6 7.073215 7 11.148881 Name: errors, dtype: float64 2020-02-01 23:54:45.972: INFO @decay_lr : LR updated to `4.056529e-05` 2020-02-01 23:54:45.973: INFO @log_profile : T train: 121.586559 2020-02-01 23:54:45.973: INFO @log_profile : T valid: 5.403939 2020-02-01 23:54:45.973: INFO @log_profile : T read data: 2.917498 2020-02-01 23:54:45.973: INFO @log_profile : T hooks: 10.029024 2020-02-01 23:54:45.973: INFO @main_loop : Epoch 180 done 2020-02-01 23:54:45.973: INFO @main_loop : Training epoch 181 2020-02-01 23:56:56.895: INFO @log_variables: train loss nanmean: 0.649688 2020-02-01 23:56:56.895: INFO @log_variables: train age_loss mean: 4.816237 2020-02-01 23:56:56.895: INFO @log_variables: train gender_loss mean: 0.100487 2020-02-01 23:56:56.895: INFO @log_variables: train age_mae mean: 5.290513 2020-02-01 23:56:56.895: INFO @log_variables: train gender_accuracy mean: 0.960606 2020-02-01 23:56:56.895: INFO @log_variables: train gender_confidence/loss nanmean: 0.048632 2020-02-01 23:56:56.895: INFO @log_variables: train gender_confidence/accuracy mean: 0.867167 2020-02-01 23:56:56.895: INFO @log_variables: train age_confidence/loss mean: 0.071865 2020-02-01 23:56:56.895: INFO @log_variables: train age_confidence/accuracy mean: 0.615464 2020-02-01 23:56:56.895: INFO @log_variables: valid loss nanmean: 0.840113 2020-02-01 23:56:56.895: INFO @log_variables: valid age_loss mean: 5.784145 2020-02-01 23:56:56.895: INFO @log_variables: valid gender_loss mean: 0.208270 2020-02-01 23:56:56.895: INFO @log_variables: valid age_mae mean: 6.264158 2020-02-01 23:56:56.895: INFO @log_variables: valid gender_accuracy mean: 0.922898 2020-02-01 23:56:56.896: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054254 2020-02-01 23:56:56.896: INFO @log_variables: valid gender_confidence/accuracy mean: 0.874993 2020-02-01 23:56:56.896: INFO @log_variables: valid age_confidence/loss mean: 0.070691 2020-02-01 23:56:56.896: INFO @log_variables: valid age_confidence/accuracy mean: 0.568683 2020-02-01 23:56:56.896: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:56:56.903: INFO @metrics_hook: train age_mae: 5.291 +-0.031 (110372) 2020-02-01 23:56:56.910: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-01 23:56:59.649: INFO @metrics_hook: valid age_mae: 6.264 +-0.090 (17639) 2020-02-01 23:56:59.650: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-01 23:57:01.346: INFO @decay_lr : LR updated to `4.036246e-05` 2020-02-01 23:57:01.348: INFO @log_profile : T train: 121.806340 2020-02-01 23:57:01.348: INFO @log_profile : T valid: 5.548384 2020-02-01 23:57:01.348: INFO @log_profile : T read data: 2.855758 2020-02-01 23:57:01.348: INFO @log_profile : T hooks: 5.085400 2020-02-01 23:57:01.348: INFO @main_loop : Epoch 181 done 2020-02-01 23:57:01.348: INFO @main_loop : Training epoch 182 2020-02-01 23:59:11.879: INFO @log_variables: train loss nanmean: 0.646752 2020-02-01 23:59:11.879: INFO @log_variables: train age_loss mean: 4.793801 2020-02-01 23:59:11.879: INFO @log_variables: train gender_loss mean: 0.099543 2020-02-01 23:59:11.879: INFO @log_variables: train age_mae mean: 5.267966 2020-02-01 23:59:11.879: INFO @log_variables: train gender_accuracy mean: 0.960476 2020-02-01 23:59:11.879: INFO @log_variables: train gender_confidence/loss nanmean: 0.048492 2020-02-01 23:59:11.879: INFO @log_variables: train gender_confidence/accuracy mean: 0.868191 2020-02-01 23:59:11.879: INFO @log_variables: train age_confidence/loss mean: 0.071966 2020-02-01 23:59:11.879: INFO @log_variables: train age_confidence/accuracy mean: 0.613399 2020-02-01 23:59:11.880: INFO @log_variables: valid loss nanmean: 0.837045 2020-02-01 23:59:11.880: INFO @log_variables: valid age_loss mean: 5.774107 2020-02-01 23:59:11.880: INFO @log_variables: valid gender_loss mean: 0.207050 2020-02-01 23:59:11.880: INFO @log_variables: valid age_mae mean: 6.253519 2020-02-01 23:59:11.880: INFO @log_variables: valid gender_accuracy mean: 0.926130 2020-02-01 23:59:11.880: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053830 2020-02-01 23:59:11.880: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875106 2020-02-01 23:59:11.880: INFO @log_variables: valid age_confidence/loss mean: 0.070069 2020-02-01 23:59:11.880: INFO @log_variables: valid age_confidence/accuracy mean: 0.557628 2020-02-01 23:59:11.880: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-01 23:59:11.887: INFO @metrics_hook: train age_mae: 5.268 +-0.030 (110592) 2020-02-01 23:59:11.894: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110592) 2020-02-01 23:59:14.655: INFO @metrics_hook: valid age_mae: 6.254 +-0.088 (17639) 2020-02-01 23:59:14.657: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-01 23:59:16.128: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:59:16.128: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-01 23:59:16.128: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-01 23:59:16.128: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-01 23:59:16.253: INFO @evaluate_confidence: Previous accuracy would be: 96.05 2020-02-01 23:59:16.253: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-01 23:59:16.318: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.4, 98.45, 98.5, 98.55, 98.6, 98.65, 98.7, 98.74, 98.78, 98.81, 98.86, 98.9, 98.93, 98.96, 99.0, 99.02, 99.05, 99.08, 99.1, 99.14, 99.18, 99.21, 99.24, 99.27, 99.29, 99.31, 99.34, 99.36, 99.39, 99.4, 99.42, 99.45, 99.47, 99.48, 99.49, 99.52, 99.53, 99.55, 99.57, 99.6, 99.62, 99.63, 99.66, 99.67, 99.69, 99.7, 99.72, 99.74, 99.74, 99.76, 99.77, 99.78] 2020-02-01 23:59:16.318: INFO @evaluate_confidence: Dropped ratios are: [9.23, 9.6, 9.94, 10.29, 10.67, 11.06, 11.44, 11.81, 12.15, 12.48, 12.83, 13.18, 13.53, 13.87, 14.21, 14.55, 14.9, 15.25, 15.62, 15.99, 16.4, 16.76, 17.15, 17.52, 17.88, 18.26, 18.7, 19.13, 19.56, 19.97, 20.34, 20.74, 21.19, 21.62, 22.09, 22.56, 23.0, 23.46, 23.94, 24.46, 24.96, 25.51, 26.01, 26.54, 27.11, 27.71, 28.32, 28.96, 29.59, 30.26, 30.97, 31.74] 2020-02-01 23:59:16.366: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:59:16.366: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-01 23:59:16.366: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-01 23:59:16.366: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-01 23:59:16.504: INFO @evaluate_confidence: Previous accuracy would be: 60.53 2020-02-01 23:59:16.504: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-01 23:59:16.519: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.48, 69.12, 69.81, 70.54, 71.29, 72.22, 73.02] 2020-02-01 23:59:16.519: INFO @evaluate_confidence: Dropped ratios are: [42.09, 45.21, 48.28, 51.42, 54.45, 57.43, 60.2] 2020-02-01 23:59:16.527: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-01 23:59:16.527: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.87 +- 0.22 2020-02-01 23:59:16.527: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.26 2020-02-01 23:59:16.527: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-01 23:59:16.635: INFO @evaluate_confidence: Previous accuracy would be: 92.61 2020-02-01 23:59:16.636: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86, 0.87] 2020-02-01 23:59:16.645: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.55, 96.65, 96.73, 96.82, 96.86, 96.93, 96.99, 97.08, 97.15, 97.18, 97.26, 97.29, 97.36, 97.39, 97.46, 97.52, 97.58, 97.63, 97.71, 97.78, 97.88, 97.9, 97.92, 97.98, 98.05, 98.09, 98.14, 98.19, 98.26, 98.27, 98.32, 98.35, 98.4, 98.47, 98.53, 98.58, 98.62, 98.68, 98.71, 98.75, 98.8, 98.85, 98.87] 2020-02-01 23:59:16.645: INFO @evaluate_confidence: Dropped ratios are: [12.97, 13.38, 13.73, 14.0, 14.29, 14.63, 14.94, 15.27, 15.53, 15.75, 16.13, 16.54, 16.92, 17.25, 17.69, 18.02, 18.41, 18.78, 19.12, 19.53, 19.89, 20.22, 20.59, 20.93, 21.37, 21.9, 22.28, 22.75, 23.24, 23.74, 24.17, 24.72, 25.2, 25.7, 26.19, 26.76, 27.3, 27.91, 28.52, 29.33, 29.99, 30.72, 31.57] 2020-02-01 23:59:16.652: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-01 23:59:16.652: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.10 2020-02-01 23:59:16.652: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-01 23:59:16.653: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-01 23:59:16.780: INFO @evaluate_confidence: Previous accuracy would be: 52.49 2020-02-01 23:59:16.780: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-01 23:59:16.781: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.81, 58.38, 59.16] 2020-02-01 23:59:16.782: INFO @evaluate_confidence: Dropped ratios are: [47.21, 51.97, 57.32] 2020-02-01 23:59:16.834: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-01 23:59:17.515: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:59:17.599: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-01 23:59:18.058: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-01 23:59:18.323: INFO @decay_lr : LR updated to `4.016065e-05` 2020-02-01 23:59:18.325: INFO @log_profile : T train: 122.435403 2020-02-01 23:59:18.325: INFO @log_profile : T valid: 5.496503 2020-02-01 23:59:18.325: INFO @log_profile : T read data: 1.906930 2020-02-01 23:59:18.325: INFO @log_profile : T hooks: 7.060114 2020-02-01 23:59:18.325: INFO @main_loop : Epoch 182 done 2020-02-01 23:59:18.325: INFO @main_loop : Training epoch 183 2020-02-02 00:01:29.197: INFO @log_variables: train loss nanmean: 0.648768 2020-02-02 00:01:29.197: INFO @log_variables: train age_loss mean: 4.785089 2020-02-02 00:01:29.197: INFO @log_variables: train gender_loss mean: 0.100424 2020-02-02 00:01:29.197: INFO @log_variables: train age_mae mean: 5.259109 2020-02-02 00:01:29.197: INFO @log_variables: train gender_accuracy mean: 0.961603 2020-02-02 00:01:29.197: INFO @log_variables: train gender_confidence/loss nanmean: 0.050368 2020-02-02 00:01:29.197: INFO @log_variables: train gender_confidence/accuracy mean: 0.863281 2020-02-02 00:01:29.197: INFO @log_variables: train age_confidence/loss mean: 0.072098 2020-02-02 00:01:29.197: INFO @log_variables: train age_confidence/accuracy mean: 0.614141 2020-02-02 00:01:29.197: INFO @log_variables: valid loss nanmean: 0.841660 2020-02-02 00:01:29.197: INFO @log_variables: valid age_loss mean: 5.766556 2020-02-02 00:01:29.197: INFO @log_variables: valid gender_loss mean: 0.212528 2020-02-02 00:01:29.197: INFO @log_variables: valid age_mae mean: 6.246749 2020-02-02 00:01:29.197: INFO @log_variables: valid gender_accuracy mean: 0.923352 2020-02-02 00:01:29.198: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054054 2020-02-02 00:01:29.198: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873859 2020-02-02 00:01:29.198: INFO @log_variables: valid age_confidence/loss mean: 0.070167 2020-02-02 00:01:29.198: INFO @log_variables: valid age_confidence/accuracy mean: 0.557288 2020-02-02 00:01:29.198: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:01:29.207: INFO @metrics_hook: train age_mae: 5.259 +-0.031 (110372) 2020-02-02 00:01:29.215: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110372) 2020-02-02 00:01:32.076: INFO @metrics_hook: valid age_mae: 6.247 +-0.089 (17639) 2020-02-02 00:01:32.077: INFO @metrics_hook: valid gender_accuracy: 0.923 +-0.004 (17639) 2020-02-02 00:01:33.851: INFO @decay_lr : LR updated to `3.9959847e-05` 2020-02-02 00:01:33.852: INFO @log_profile : T train: 121.829299 2020-02-02 00:01:33.852: INFO @log_profile : T valid: 5.435484 2020-02-02 00:01:33.852: INFO @log_profile : T read data: 2.897561 2020-02-02 00:01:33.852: INFO @log_profile : T hooks: 5.289110 2020-02-02 00:01:33.853: INFO @main_loop : Epoch 183 done 2020-02-02 00:01:33.853: INFO @main_loop : Training epoch 184 2020-02-02 00:03:44.059: INFO @log_variables: train loss nanmean: 0.649853 2020-02-02 00:03:44.059: INFO @log_variables: train age_loss mean: 4.793433 2020-02-02 00:03:44.059: INFO @log_variables: train gender_loss mean: 0.102159 2020-02-02 00:03:44.059: INFO @log_variables: train age_mae mean: 5.267514 2020-02-02 00:03:44.059: INFO @log_variables: train gender_accuracy mean: 0.960024 2020-02-02 00:03:44.059: INFO @log_variables: train gender_confidence/loss nanmean: 0.049359 2020-02-02 00:03:44.060: INFO @log_variables: train gender_confidence/accuracy mean: 0.865008 2020-02-02 00:03:44.060: INFO @log_variables: train age_confidence/loss mean: 0.071855 2020-02-02 00:03:44.060: INFO @log_variables: train age_confidence/accuracy mean: 0.613977 2020-02-02 00:03:44.060: INFO @log_variables: valid loss nanmean: 0.852680 2020-02-02 00:03:44.060: INFO @log_variables: valid age_loss mean: 5.851999 2020-02-02 00:03:44.060: INFO @log_variables: valid gender_loss mean: 0.213523 2020-02-02 00:03:44.060: INFO @log_variables: valid age_mae mean: 6.332375 2020-02-02 00:03:44.060: INFO @log_variables: valid gender_accuracy mean: 0.924032 2020-02-02 00:03:44.060: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055915 2020-02-02 00:03:44.060: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872782 2020-02-02 00:03:44.060: INFO @log_variables: valid age_confidence/loss mean: 0.070653 2020-02-02 00:03:44.060: INFO @log_variables: valid age_confidence/accuracy mean: 0.559726 2020-02-02 00:03:44.060: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:03:44.067: INFO @metrics_hook: train age_mae: 5.268 +-0.030 (110592) 2020-02-02 00:03:44.075: INFO @metrics_hook: train gender_accuracy: 0.960 +-0.001 (110592) 2020-02-02 00:03:46.946: INFO @metrics_hook: valid age_mae: 6.332 +-0.090 (17639) 2020-02-02 00:03:46.948: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-02 00:03:48.518: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:03:48.518: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-02 00:03:48.519: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:03:48.519: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-02 00:03:48.652: INFO @evaluate_confidence: Previous accuracy would be: 96.00 2020-02-02 00:03:48.652: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-02 00:03:48.718: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.33, 98.4, 98.46, 98.52, 98.56, 98.61, 98.66, 98.72, 98.74, 98.79, 98.83, 98.88, 98.92, 98.96, 99.0, 99.03, 99.06, 99.09, 99.12, 99.15, 99.18, 99.2, 99.24, 99.26, 99.28, 99.3, 99.32, 99.35, 99.38, 99.41, 99.43, 99.46, 99.48, 99.49, 99.51, 99.54, 99.56, 99.57, 99.58, 99.6, 99.61, 99.63, 99.64, 99.66, 99.68, 99.69, 99.7, 99.71, 99.73, 99.74, 99.76] 2020-02-02 00:03:48.719: INFO @evaluate_confidence: Dropped ratios are: [9.28, 9.71, 10.11, 10.49, 10.9, 11.25, 11.63, 11.98, 12.36, 12.74, 13.09, 13.45, 13.82, 14.18, 14.56, 14.94, 15.28, 15.66, 16.02, 16.42, 16.77, 17.14, 17.49, 17.86, 18.23, 18.65, 19.01, 19.41, 19.82, 20.23, 20.66, 21.09, 21.56, 22.0, 22.42, 22.9, 23.36, 23.83, 24.33, 24.83, 25.32, 25.83, 26.37, 26.93, 27.52, 28.13, 28.72, 29.37, 30.08, 30.79, 31.52] 2020-02-02 00:03:48.770: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:03:48.770: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:03:48.770: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:03:48.770: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:03:48.917: INFO @evaluate_confidence: Previous accuracy would be: 60.29 2020-02-02 00:03:48.917: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:03:48.932: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.5, 69.15, 69.87, 70.61, 71.39, 72.04, 72.92] 2020-02-02 00:03:48.932: INFO @evaluate_confidence: Dropped ratios are: [42.61, 45.82, 49.03, 51.99, 54.92, 57.73, 60.4] 2020-02-02 00:03:48.940: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:03:48.940: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-02 00:03:48.940: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-02 00:03:48.940: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-02 00:03:49.051: INFO @evaluate_confidence: Previous accuracy would be: 92.40 2020-02-02 00:03:49.051: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:03:49.060: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.37, 96.46, 96.55, 96.61, 96.71, 96.84, 96.92, 97.01, 97.1, 97.17, 97.22, 97.25, 97.31, 97.4, 97.48, 97.52, 97.55, 97.59, 97.62, 97.66, 97.76, 97.81, 97.86, 97.92, 97.95, 97.98, 98.08, 98.12, 98.19, 98.26, 98.33, 98.4, 98.45, 98.49, 98.53, 98.59, 98.63, 98.65, 98.7, 98.73, 98.77, 98.81] 2020-02-02 00:03:49.061: INFO @evaluate_confidence: Dropped ratios are: [13.07, 13.43, 13.76, 14.16, 14.51, 14.94, 15.23, 15.66, 16.13, 16.54, 16.88, 17.26, 17.73, 18.14, 18.62, 18.99, 19.41, 19.76, 20.09, 20.52, 20.96, 21.33, 21.74, 22.21, 22.6, 22.98, 23.46, 23.95, 24.45, 24.85, 25.34, 25.83, 26.31, 27.04, 27.55, 28.18, 28.78, 29.5, 30.18, 30.85, 31.52, 32.34] 2020-02-02 00:03:49.068: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:03:49.068: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.10 2020-02-02 00:03:49.069: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.10 2020-02-02 00:03:49.069: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.10 2020-02-02 00:03:49.201: INFO @evaluate_confidence: Previous accuracy would be: 52.70 2020-02-02 00:03:49.202: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-02 00:03:49.203: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.45, 57.84, 58.75, 59.45, 60.09] 2020-02-02 00:03:49.203: INFO @evaluate_confidence: Dropped ratios are: [44.3, 49.43, 54.36, 59.4, 64.11] 2020-02-02 00:03:49.256: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:03:50.026: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:03:50.114: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:03:50.623: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:03:50.907: INFO @decay_lr : LR updated to `3.976005e-05` 2020-02-02 00:03:50.908: INFO @log_profile : T train: 121.898166 2020-02-02 00:03:50.908: INFO @log_profile : T valid: 5.565748 2020-02-02 00:03:50.908: INFO @log_profile : T read data: 2.059307 2020-02-02 00:03:50.908: INFO @log_profile : T hooks: 7.455561 2020-02-02 00:03:50.908: INFO @main_loop : Epoch 184 done 2020-02-02 00:03:50.908: INFO @main_loop : Training epoch 185 2020-02-02 00:06:02.020: INFO @log_variables: train loss nanmean: 0.647800 2020-02-02 00:06:02.020: INFO @log_variables: train age_loss mean: 4.794311 2020-02-02 00:06:02.020: INFO @log_variables: train gender_loss mean: 0.100014 2020-02-02 00:06:02.020: INFO @log_variables: train age_mae mean: 5.268278 2020-02-02 00:06:02.020: INFO @log_variables: train gender_accuracy mean: 0.961168 2020-02-02 00:06:02.020: INFO @log_variables: train gender_confidence/loss nanmean: 0.049236 2020-02-02 00:06:02.020: INFO @log_variables: train gender_confidence/accuracy mean: 0.865174 2020-02-02 00:06:02.020: INFO @log_variables: train age_confidence/loss mean: 0.071796 2020-02-02 00:06:02.020: INFO @log_variables: train age_confidence/accuracy mean: 0.613987 2020-02-02 00:06:02.020: INFO @log_variables: valid loss nanmean: 0.850545 2020-02-02 00:06:02.020: INFO @log_variables: valid age_loss mean: 5.762531 2020-02-02 00:06:02.020: INFO @log_variables: valid gender_loss mean: 0.218053 2020-02-02 00:06:02.020: INFO @log_variables: valid age_mae mean: 6.243148 2020-02-02 00:06:02.020: INFO @log_variables: valid gender_accuracy mean: 0.926016 2020-02-02 00:06:02.021: INFO @log_variables: valid gender_confidence/loss nanmean: 0.058593 2020-02-02 00:06:02.021: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875390 2020-02-02 00:06:02.021: INFO @log_variables: valid age_confidence/loss mean: 0.069856 2020-02-02 00:06:02.021: INFO @log_variables: valid age_confidence/accuracy mean: 0.554567 2020-02-02 00:06:02.021: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:06:02.029: INFO @metrics_hook: train age_mae: 5.268 +-0.030 (110372) 2020-02-02 00:06:02.036: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:06:04.951: INFO @metrics_hook: valid age_mae: 6.243 +-0.088 (17639) 2020-02-02 00:06:04.952: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-02 00:06:06.671: INFO @decay_lr : LR updated to `3.9561248e-05` 2020-02-02 00:06:06.672: INFO @log_profile : T train: 121.889165 2020-02-02 00:06:06.672: INFO @log_profile : T valid: 5.527672 2020-02-02 00:06:06.673: INFO @log_profile : T read data: 3.024655 2020-02-02 00:06:06.673: INFO @log_profile : T hooks: 5.243245 2020-02-02 00:06:06.673: INFO @main_loop : Epoch 185 done 2020-02-02 00:06:06.673: INFO @main_loop : Training epoch 186 2020-02-02 00:08:18.995: INFO @log_variables: train loss nanmean: 0.647107 2020-02-02 00:08:18.996: INFO @log_variables: train age_loss mean: 4.761136 2020-02-02 00:08:18.996: INFO @log_variables: train gender_loss mean: 0.101570 2020-02-02 00:08:18.996: INFO @log_variables: train age_mae mean: 5.235362 2020-02-02 00:08:18.996: INFO @log_variables: train gender_accuracy mean: 0.960660 2020-02-02 00:08:18.996: INFO @log_variables: train gender_confidence/loss nanmean: 0.049898 2020-02-02 00:08:18.996: INFO @log_variables: train gender_confidence/accuracy mean: 0.864440 2020-02-02 00:08:18.996: INFO @log_variables: train age_confidence/loss mean: 0.072042 2020-02-02 00:08:18.996: INFO @log_variables: train age_confidence/accuracy mean: 0.614042 2020-02-02 00:08:18.996: INFO @log_variables: valid loss nanmean: 0.838691 2020-02-02 00:08:18.996: INFO @log_variables: valid age_loss mean: 5.769230 2020-02-02 00:08:18.996: INFO @log_variables: valid gender_loss mean: 0.208142 2020-02-02 00:08:18.996: INFO @log_variables: valid age_mae mean: 6.249853 2020-02-02 00:08:18.996: INFO @log_variables: valid gender_accuracy mean: 0.925733 2020-02-02 00:08:18.996: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054348 2020-02-02 00:08:18.996: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867509 2020-02-02 00:08:18.996: INFO @log_variables: valid age_confidence/loss mean: 0.070648 2020-02-02 00:08:18.997: INFO @log_variables: valid age_confidence/accuracy mean: 0.565792 2020-02-02 00:08:18.997: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:08:19.004: INFO @metrics_hook: train age_mae: 5.235 +-0.030 (110372) 2020-02-02 00:08:19.012: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:08:21.919: INFO @metrics_hook: valid age_mae: 6.250 +-0.090 (17639) 2020-02-02 00:08:21.921: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-02 00:08:23.450: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:08:23.450: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:08:23.450: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.21 2020-02-02 00:08:23.450: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:08:23.584: INFO @evaluate_confidence: Previous accuracy would be: 96.07 2020-02-02 00:08:23.584: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:08:23.649: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.36, 98.4, 98.47, 98.51, 98.56, 98.61, 98.66, 98.71, 98.75, 98.79, 98.82, 98.88, 98.91, 98.95, 98.99, 99.03, 99.06, 99.1, 99.14, 99.17, 99.2, 99.22, 99.26, 99.28, 99.31, 99.33, 99.36, 99.39, 99.41, 99.43, 99.44, 99.46, 99.49, 99.51, 99.53, 99.55, 99.56, 99.57, 99.58, 99.6, 99.62, 99.63, 99.65, 99.67, 99.69, 99.7, 99.71, 99.73, 99.74, 99.75, 99.76] 2020-02-02 00:08:23.649: INFO @evaluate_confidence: Dropped ratios are: [9.63, 10.03, 10.4, 10.8, 11.15, 11.51, 11.89, 12.24, 12.61, 12.97, 13.33, 13.74, 14.09, 14.47, 14.85, 15.24, 15.58, 15.98, 16.37, 16.76, 17.14, 17.48, 17.87, 18.25, 18.65, 19.05, 19.46, 19.85, 20.26, 20.65, 21.04, 21.48, 21.93, 22.38, 22.85, 23.3, 23.76, 24.25, 24.74, 25.25, 25.76, 26.27, 26.82, 27.4, 27.96, 28.56, 29.18, 29.84, 30.51, 31.21, 31.95] 2020-02-02 00:08:23.698: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:08:23.699: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:08:23.699: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:08:23.699: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:08:23.843: INFO @evaluate_confidence: Previous accuracy would be: 60.80 2020-02-02 00:08:23.843: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:08:23.859: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.7, 69.4, 70.22, 70.9, 71.65, 72.51, 73.32] 2020-02-02 00:08:23.859: INFO @evaluate_confidence: Dropped ratios are: [41.83, 44.94, 47.98, 51.02, 54.02, 56.88, 59.64] 2020-02-02 00:08:23.867: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:08:23.867: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-02 00:08:23.867: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.25 2020-02-02 00:08:23.867: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:08:23.977: INFO @evaluate_confidence: Previous accuracy would be: 92.57 2020-02-02 00:08:23.978: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-02 00:08:23.987: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.48, 96.61, 96.72, 96.78, 96.84, 96.93, 97.02, 97.09, 97.19, 97.26, 97.33, 97.4, 97.46, 97.49, 97.55, 97.61, 97.62, 97.68, 97.71, 97.76, 97.83, 97.89, 97.98, 98.04, 98.11, 98.16, 98.2, 98.24, 98.27, 98.3, 98.35, 98.4, 98.45, 98.49, 98.54, 98.55, 98.62, 98.68, 98.75, 98.79, 98.83, 98.86, 98.92] 2020-02-02 00:08:23.987: INFO @evaluate_confidence: Dropped ratios are: [13.23, 13.63, 14.01, 14.36, 14.73, 15.16, 15.44, 15.78, 16.15, 16.46, 16.8, 17.14, 17.49, 17.86, 18.22, 18.57, 18.89, 19.25, 19.63, 19.99, 20.36, 20.79, 21.31, 21.75, 22.16, 22.56, 23.05, 23.47, 23.98, 24.42, 24.92, 25.43, 25.98, 26.4, 26.94, 27.51, 28.11, 28.73, 29.43, 30.06, 30.81, 31.63, 32.53] 2020-02-02 00:08:23.995: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:08:23.995: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.11 2020-02-02 00:08:23.995: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-02 00:08:23.995: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-02 00:08:24.129: INFO @evaluate_confidence: Previous accuracy would be: 53.12 2020-02-02 00:08:24.129: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-02 00:08:24.131: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.86, 59.32, 59.8, 60.43, 60.89] 2020-02-02 00:08:24.131: INFO @evaluate_confidence: Dropped ratios are: [45.26, 50.0, 54.84, 59.46, 64.05] 2020-02-02 00:08:24.187: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:08:36.784: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:08:36.871: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:08:37.371: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:08:37.646: INFO @decay_lr : LR updated to `3.936344e-05` 2020-02-02 00:08:37.647: INFO @log_profile : T train: 123.121617 2020-02-02 00:08:37.647: INFO @log_profile : T valid: 5.487468 2020-02-02 00:08:37.647: INFO @log_profile : T read data: 3.041873 2020-02-02 00:08:37.647: INFO @log_profile : T hooks: 19.247132 2020-02-02 00:08:37.647: INFO @main_loop : Epoch 186 done 2020-02-02 00:08:37.647: INFO @main_loop : Training epoch 187 2020-02-02 00:10:56.185: INFO @log_variables: train loss nanmean: 0.648304 2020-02-02 00:10:56.185: INFO @log_variables: train age_loss mean: 4.775784 2020-02-02 00:10:56.185: INFO @log_variables: train gender_loss mean: 0.101258 2020-02-02 00:10:56.185: INFO @log_variables: train age_mae mean: 5.250235 2020-02-02 00:10:56.185: INFO @log_variables: train gender_accuracy mean: 0.960919 2020-02-02 00:10:56.185: INFO @log_variables: train gender_confidence/loss nanmean: 0.050097 2020-02-02 00:10:56.185: INFO @log_variables: train gender_confidence/accuracy mean: 0.862015 2020-02-02 00:10:56.185: INFO @log_variables: train age_confidence/loss mean: 0.071993 2020-02-02 00:10:56.185: INFO @log_variables: train age_confidence/accuracy mean: 0.615903 2020-02-02 00:10:56.185: INFO @log_variables: valid loss nanmean: 0.844875 2020-02-02 00:10:56.185: INFO @log_variables: valid age_loss mean: 5.832224 2020-02-02 00:10:56.186: INFO @log_variables: valid gender_loss mean: 0.208561 2020-02-02 00:10:56.186: INFO @log_variables: valid age_mae mean: 6.311889 2020-02-02 00:10:56.186: INFO @log_variables: valid gender_accuracy mean: 0.924032 2020-02-02 00:10:56.186: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054405 2020-02-02 00:10:56.186: INFO @log_variables: valid gender_confidence/accuracy mean: 0.867963 2020-02-02 00:10:56.186: INFO @log_variables: valid age_confidence/loss mean: 0.070667 2020-02-02 00:10:56.186: INFO @log_variables: valid age_confidence/accuracy mean: 0.554397 2020-02-02 00:10:56.186: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:10:56.193: INFO @metrics_hook: train age_mae: 5.250 +-0.030 (110592) 2020-02-02 00:10:56.201: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110592) 2020-02-02 00:10:59.120: INFO @metrics_hook: valid age_mae: 6.312 +-0.091 (17639) 2020-02-02 00:10:59.121: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-02 00:11:00.871: INFO @decay_lr : LR updated to `3.916662e-05` 2020-02-02 00:11:00.872: INFO @log_profile : T train: 130.496329 2020-02-02 00:11:00.872: INFO @log_profile : T valid: 5.494264 2020-02-02 00:11:00.872: INFO @log_profile : T read data: 1.894623 2020-02-02 00:11:00.872: INFO @log_profile : T hooks: 5.263684 2020-02-02 00:11:00.872: INFO @main_loop : Epoch 187 done 2020-02-02 00:11:00.872: INFO @main_loop : Training epoch 188 2020-02-02 00:13:11.754: INFO @log_variables: train loss nanmean: 0.647274 2020-02-02 00:13:11.754: INFO @log_variables: train age_loss mean: 4.782397 2020-02-02 00:13:11.754: INFO @log_variables: train gender_loss mean: 0.099901 2020-02-02 00:13:11.754: INFO @log_variables: train age_mae mean: 5.256428 2020-02-02 00:13:11.754: INFO @log_variables: train gender_accuracy mean: 0.961838 2020-02-02 00:13:11.754: INFO @log_variables: train gender_confidence/loss nanmean: 0.049956 2020-02-02 00:13:11.754: INFO @log_variables: train gender_confidence/accuracy mean: 0.864875 2020-02-02 00:13:11.754: INFO @log_variables: train age_confidence/loss mean: 0.071736 2020-02-02 00:13:11.754: INFO @log_variables: train age_confidence/accuracy mean: 0.616986 2020-02-02 00:13:11.754: INFO @log_variables: valid loss nanmean: 0.847878 2020-02-02 00:13:11.754: INFO @log_variables: valid age_loss mean: 5.807620 2020-02-02 00:13:11.754: INFO @log_variables: valid gender_loss mean: 0.214362 2020-02-02 00:13:11.754: INFO @log_variables: valid age_mae mean: 6.288129 2020-02-02 00:13:11.755: INFO @log_variables: valid gender_accuracy mean: 0.924939 2020-02-02 00:13:11.755: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054907 2020-02-02 00:13:11.755: INFO @log_variables: valid gender_confidence/accuracy mean: 0.864618 2020-02-02 00:13:11.755: INFO @log_variables: valid age_confidence/loss mean: 0.070131 2020-02-02 00:13:11.755: INFO @log_variables: valid age_confidence/accuracy mean: 0.555757 2020-02-02 00:13:11.755: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:13:11.763: INFO @metrics_hook: train age_mae: 5.256 +-0.031 (110372) 2020-02-02 00:13:11.770: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110372) 2020-02-02 00:13:14.647: INFO @metrics_hook: valid age_mae: 6.288 +-0.089 (17639) 2020-02-02 00:13:14.648: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-02 00:13:16.195: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:13:16.195: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:13:16.195: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.33 +- 0.22 2020-02-02 00:13:16.195: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:13:16.326: INFO @evaluate_confidence: Previous accuracy would be: 96.18 2020-02-02 00:13:16.327: INFO @evaluate_confidence: Possible optimal thresholds are: [0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:13:16.390: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.42, 98.46, 98.51, 98.57, 98.62, 98.67, 98.72, 98.76, 98.8, 98.84, 98.87, 98.93, 98.96, 99.0, 99.03, 99.06, 99.09, 99.12, 99.14, 99.17, 99.2, 99.22, 99.25, 99.27, 99.3, 99.32, 99.34, 99.35, 99.37, 99.4, 99.42, 99.44, 99.46, 99.47, 99.49, 99.51, 99.53, 99.56, 99.57, 99.59, 99.62, 99.63, 99.63, 99.65, 99.66, 99.68, 99.7, 99.71, 99.74, 99.75, 99.76] 2020-02-02 00:13:16.391: INFO @evaluate_confidence: Dropped ratios are: [9.59, 9.95, 10.32, 10.7, 11.08, 11.45, 11.82, 12.21, 12.56, 12.94, 13.31, 13.7, 14.05, 14.42, 14.77, 15.14, 15.49, 15.84, 16.19, 16.57, 16.96, 17.34, 17.75, 18.12, 18.5, 18.9, 19.29, 19.69, 20.08, 20.53, 20.93, 21.37, 21.81, 22.23, 22.67, 23.16, 23.62, 24.12, 24.61, 25.1, 25.61, 26.14, 26.69, 27.28, 27.88, 28.5, 29.19, 29.87, 30.56, 31.3, 32.08] 2020-02-02 00:13:16.442: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:13:16.443: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:13:16.443: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.12 2020-02-02 00:13:16.443: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:13:16.586: INFO @evaluate_confidence: Previous accuracy would be: 60.50 2020-02-02 00:13:16.586: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:13:16.603: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.38, 69.08, 69.75, 70.48, 71.14, 71.97, 72.77, 73.6] 2020-02-02 00:13:16.603: INFO @evaluate_confidence: Dropped ratios are: [39.63, 42.76, 45.79, 48.88, 51.88, 54.78, 57.69, 60.39] 2020-02-02 00:13:16.611: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:13:16.611: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.85 +- 0.23 2020-02-02 00:13:16.611: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.25 2020-02-02 00:13:16.611: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:13:16.730: INFO @evaluate_confidence: Previous accuracy would be: 92.49 2020-02-02 00:13:16.731: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85] 2020-02-02 00:13:16.744: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.56, 96.61, 96.66, 96.78, 96.86, 96.95, 97.05, 97.16, 97.24, 97.28, 97.32, 97.38, 97.45, 97.5, 97.56, 97.6, 97.69, 97.75, 97.78, 97.83, 97.91, 97.92, 97.94, 97.99, 98.02, 98.08, 98.11, 98.15, 98.19, 98.2, 98.24, 98.32, 98.37, 98.42, 98.47, 98.52, 98.55, 98.56, 98.6, 98.66, 98.71, 98.76] 2020-02-02 00:13:16.744: INFO @evaluate_confidence: Dropped ratios are: [13.92, 14.25, 14.58, 15.02, 15.42, 15.74, 16.09, 16.49, 16.9, 17.29, 17.65, 18.05, 18.47, 18.75, 19.1, 19.45, 19.82, 20.18, 20.57, 20.95, 21.35, 21.7, 22.06, 22.46, 22.86, 23.33, 23.78, 24.35, 24.83, 25.15, 25.66, 26.18, 26.74, 27.43, 27.97, 28.56, 29.16, 29.7, 30.54, 31.15, 31.96, 32.8] 2020-02-02 00:13:16.756: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:13:16.756: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-02 00:13:16.757: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-02 00:13:16.757: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-02 00:13:16.898: INFO @evaluate_confidence: Previous accuracy would be: 52.53 2020-02-02 00:13:16.898: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-02 00:13:16.899: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.57, 57.61, 58.24] 2020-02-02 00:13:16.899: INFO @evaluate_confidence: Dropped ratios are: [46.48, 51.11, 55.54] 2020-02-02 00:13:16.953: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:13:17.690: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:13:17.780: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:13:18.274: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:13:18.557: INFO @decay_lr : LR updated to `3.897079e-05` 2020-02-02 00:13:18.559: INFO @log_profile : T train: 121.655182 2020-02-02 00:13:18.559: INFO @log_profile : T valid: 5.507358 2020-02-02 00:13:18.559: INFO @log_profile : T read data: 3.042120 2020-02-02 00:13:18.559: INFO @log_profile : T hooks: 7.405802 2020-02-02 00:13:18.559: INFO @main_loop : Epoch 188 done 2020-02-02 00:13:18.559: INFO @main_loop : Training epoch 189 2020-02-02 00:15:29.662: INFO @log_variables: train loss nanmean: 0.644001 2020-02-02 00:15:29.663: INFO @log_variables: train age_loss mean: 4.763476 2020-02-02 00:15:29.663: INFO @log_variables: train gender_loss mean: 0.099695 2020-02-02 00:15:29.663: INFO @log_variables: train age_mae mean: 5.238045 2020-02-02 00:15:29.663: INFO @log_variables: train gender_accuracy mean: 0.961068 2020-02-02 00:15:29.663: INFO @log_variables: train gender_confidence/loss nanmean: 0.048356 2020-02-02 00:15:29.663: INFO @log_variables: train gender_confidence/accuracy mean: 0.866234 2020-02-02 00:15:29.663: INFO @log_variables: train age_confidence/loss mean: 0.071970 2020-02-02 00:15:29.663: INFO @log_variables: train age_confidence/accuracy mean: 0.615065 2020-02-02 00:15:29.663: INFO @log_variables: valid loss nanmean: 0.841801 2020-02-02 00:15:29.663: INFO @log_variables: valid age_loss mean: 5.747677 2020-02-02 00:15:29.663: INFO @log_variables: valid gender_loss mean: 0.212149 2020-02-02 00:15:29.663: INFO @log_variables: valid age_mae mean: 6.227728 2020-02-02 00:15:29.663: INFO @log_variables: valid gender_accuracy mean: 0.925676 2020-02-02 00:15:29.663: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056052 2020-02-02 00:15:29.663: INFO @log_variables: valid gender_confidence/accuracy mean: 0.877884 2020-02-02 00:15:29.663: INFO @log_variables: valid age_confidence/loss mean: 0.070370 2020-02-02 00:15:29.664: INFO @log_variables: valid age_confidence/accuracy mean: 0.563014 2020-02-02 00:15:29.664: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:15:29.671: INFO @metrics_hook: train age_mae: 5.238 +-0.030 (110372) 2020-02-02 00:15:29.678: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:15:32.551: INFO @metrics_hook: valid age_mae: 6.228 +-0.088 (17639) 2020-02-02 00:15:32.552: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-02 00:15:34.381: INFO @decay_lr : LR updated to `3.8775936e-05` 2020-02-02 00:15:34.382: INFO @log_profile : T train: 121.879985 2020-02-02 00:15:34.382: INFO @log_profile : T valid: 5.445394 2020-02-02 00:15:34.382: INFO @log_profile : T read data: 3.092811 2020-02-02 00:15:34.383: INFO @log_profile : T hooks: 5.328729 2020-02-02 00:15:34.383: INFO @main_loop : Epoch 189 done 2020-02-02 00:15:34.383: INFO @main_loop : Training epoch 190 2020-02-02 00:17:46.185: INFO @log_variables: train loss nanmean: 0.646253 2020-02-02 00:17:46.185: INFO @log_variables: train age_loss mean: 4.784061 2020-02-02 00:17:46.185: INFO @log_variables: train gender_loss mean: 0.099251 2020-02-02 00:17:46.185: INFO @log_variables: train age_mae mean: 5.258187 2020-02-02 00:17:46.185: INFO @log_variables: train gender_accuracy mean: 0.961760 2020-02-02 00:17:46.185: INFO @log_variables: train gender_confidence/loss nanmean: 0.049221 2020-02-02 00:17:46.185: INFO @log_variables: train gender_confidence/accuracy mean: 0.865189 2020-02-02 00:17:46.185: INFO @log_variables: train age_confidence/loss mean: 0.071889 2020-02-02 00:17:46.185: INFO @log_variables: train age_confidence/accuracy mean: 0.615623 2020-02-02 00:17:46.185: INFO @log_variables: valid loss nanmean: 0.867150 2020-02-02 00:17:46.185: INFO @log_variables: valid age_loss mean: 5.697064 2020-02-02 00:17:46.185: INFO @log_variables: valid gender_loss mean: 0.243571 2020-02-02 00:17:46.185: INFO @log_variables: valid age_mae mean: 6.176783 2020-02-02 00:17:46.185: INFO @log_variables: valid gender_accuracy mean: 0.915528 2020-02-02 00:17:46.186: INFO @log_variables: valid gender_confidence/loss nanmean: 0.057095 2020-02-02 00:17:46.186: INFO @log_variables: valid gender_confidence/accuracy mean: 0.865185 2020-02-02 00:17:46.186: INFO @log_variables: valid age_confidence/loss mean: 0.070712 2020-02-02 00:17:46.186: INFO @log_variables: valid age_confidence/accuracy mean: 0.554680 2020-02-02 00:17:46.186: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:17:46.194: INFO @metrics_hook: train age_mae: 5.258 +-0.030 (110592) 2020-02-02 00:17:46.202: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110592) 2020-02-02 00:17:49.110: INFO @metrics_hook: valid age_mae: 6.177 +-0.088 (17639) 2020-02-02 00:17:49.111: INFO @metrics_hook: valid gender_accuracy: 0.916 +-0.004 (17639) 2020-02-02 00:17:50.687: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:17:50.687: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:17:50.687: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:17:50.688: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:17:50.816: INFO @evaluate_confidence: Previous accuracy would be: 96.18 2020-02-02 00:17:50.816: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:17:50.882: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.45, 98.51, 98.57, 98.62, 98.67, 98.72, 98.76, 98.79, 98.82, 98.87, 98.9, 98.94, 98.98, 99.01, 99.04, 99.09, 99.12, 99.14, 99.17, 99.19, 99.22, 99.25, 99.27, 99.29, 99.31, 99.32, 99.36, 99.38, 99.41, 99.43, 99.45, 99.47, 99.48, 99.5, 99.51, 99.52, 99.53, 99.55, 99.57, 99.59, 99.59, 99.6, 99.62, 99.64, 99.66, 99.67, 99.69, 99.71, 99.72, 99.73, 99.75, 99.76] 2020-02-02 00:17:50.882: INFO @evaluate_confidence: Dropped ratios are: [9.33, 9.72, 10.1, 10.46, 10.86, 11.22, 11.61, 11.98, 12.32, 12.67, 13.01, 13.37, 13.72, 14.06, 14.43, 14.81, 15.17, 15.54, 15.91, 16.27, 16.65, 17.01, 17.37, 17.75, 18.12, 18.47, 18.87, 19.26, 19.65, 20.07, 20.51, 20.89, 21.31, 21.75, 22.19, 22.62, 23.08, 23.56, 24.01, 24.5, 24.99, 25.52, 26.06, 26.65, 27.27, 27.86, 28.49, 29.12, 29.76, 30.44, 31.15, 31.91] 2020-02-02 00:17:50.934: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:17:50.934: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:17:50.935: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:17:50.935: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:17:51.080: INFO @evaluate_confidence: Previous accuracy would be: 60.46 2020-02-02 00:17:51.081: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:17:51.096: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.52, 69.22, 69.93, 70.6, 71.36, 72.22, 73.04] 2020-02-02 00:17:51.096: INFO @evaluate_confidence: Dropped ratios are: [41.77, 44.95, 48.09, 51.19, 54.18, 57.04, 59.76] 2020-02-02 00:17:51.105: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:17:51.105: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.23 2020-02-02 00:17:51.105: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.26 2020-02-02 00:17:51.105: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-02 00:17:51.212: INFO @evaluate_confidence: Previous accuracy would be: 91.55 2020-02-02 00:17:51.213: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:17:51.221: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [95.89, 96.02, 96.19, 96.25, 96.35, 96.37, 96.44, 96.54, 96.6, 96.68, 96.76, 96.84, 96.93, 96.99, 97.06, 97.08, 97.16, 97.22, 97.26, 97.31, 97.35, 97.38, 97.46, 97.54, 97.58, 97.62, 97.69, 97.71, 97.81, 97.83, 97.89, 97.95, 98.04, 98.08, 98.16, 98.24, 98.33, 98.37, 98.37, 98.4] 2020-02-02 00:17:51.221: INFO @evaluate_confidence: Dropped ratios are: [14.38, 14.79, 15.21, 15.59, 15.99, 16.4, 16.8, 17.33, 17.68, 18.08, 18.49, 18.91, 19.25, 19.62, 20.0, 20.39, 20.85, 21.23, 21.6, 22.05, 22.44, 22.87, 23.36, 23.78, 24.25, 24.8, 25.26, 25.68, 26.12, 26.63, 27.12, 27.63, 28.3, 28.77, 29.32, 29.89, 30.51, 31.18, 31.91, 32.62] 2020-02-02 00:17:51.229: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:17:51.229: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-02 00:17:51.229: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.10 2020-02-02 00:17:51.230: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-02 00:17:51.359: INFO @evaluate_confidence: Previous accuracy would be: 53.31 2020-02-02 00:17:51.360: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52] 2020-02-02 00:17:51.361: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.27, 58.64, 59.62] 2020-02-02 00:17:51.361: INFO @evaluate_confidence: Dropped ratios are: [46.88, 51.39, 56.1] 2020-02-02 00:17:51.417: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:17:52.151: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:17:52.238: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:17:52.717: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:17:52.793: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:17:53.565: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:17:53.653: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-02 00:17:53.655: INFO @evaluate_gender-age_model: groups 0 3.204780 1 3.648148 2 4.939182 3 5.397579 4 5.972923 5 5.922225 6 5.844916 7 6.417281 Name: errors, dtype: float64 2020-02-02 00:17:53.656: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:17:54.131: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:17:54.193: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-02 00:17:54.195: INFO @evaluate_gender-age_model: groups 0 6.028337 1 5.470957 2 5.527853 3 5.590772 4 7.132421 5 5.404448 6 7.539733 7 11.457036 Name: errors, dtype: float64 2020-02-02 00:17:54.389: INFO @decay_lr : LR updated to `3.8582057e-05` 2020-02-02 00:17:54.390: INFO @log_profile : T train: 121.886886 2020-02-02 00:17:54.390: INFO @log_profile : T valid: 5.394785 2020-02-02 00:17:54.390: INFO @log_profile : T read data: 2.072533 2020-02-02 00:17:54.390: INFO @log_profile : T hooks: 10.576485 2020-02-02 00:17:54.390: INFO @main_loop : Epoch 190 done 2020-02-02 00:17:54.390: INFO @main_loop : Training epoch 191 2020-02-02 00:20:05.361: INFO @log_variables: train loss nanmean: 0.645274 2020-02-02 00:20:05.361: INFO @log_variables: train age_loss mean: 4.747540 2020-02-02 00:20:05.361: INFO @log_variables: train gender_loss mean: 0.100856 2020-02-02 00:20:05.362: INFO @log_variables: train age_mae mean: 5.221957 2020-02-02 00:20:05.362: INFO @log_variables: train gender_accuracy mean: 0.960923 2020-02-02 00:20:05.362: INFO @log_variables: train gender_confidence/loss nanmean: 0.049813 2020-02-02 00:20:05.362: INFO @log_variables: train gender_confidence/accuracy mean: 0.864830 2020-02-02 00:20:05.362: INFO @log_variables: train age_confidence/loss mean: 0.072180 2020-02-02 00:20:05.362: INFO @log_variables: train age_confidence/accuracy mean: 0.612547 2020-02-02 00:20:05.362: INFO @log_variables: valid loss nanmean: 0.823241 2020-02-02 00:20:05.362: INFO @log_variables: valid age_loss mean: 5.712170 2020-02-02 00:20:05.362: INFO @log_variables: valid gender_loss mean: 0.198529 2020-02-02 00:20:05.362: INFO @log_variables: valid age_mae mean: 6.192739 2020-02-02 00:20:05.362: INFO @log_variables: valid gender_accuracy mean: 0.928341 2020-02-02 00:20:05.362: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053379 2020-02-02 00:20:05.362: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866659 2020-02-02 00:20:05.362: INFO @log_variables: valid age_confidence/loss mean: 0.070092 2020-02-02 00:20:05.362: INFO @log_variables: valid age_confidence/accuracy mean: 0.558819 2020-02-02 00:20:05.362: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:20:05.370: INFO @metrics_hook: train age_mae: 5.222 +-0.030 (110372) 2020-02-02 00:20:05.377: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:20:08.203: INFO @metrics_hook: valid age_mae: 6.193 +-0.087 (17639) 2020-02-02 00:20:08.204: INFO @metrics_hook: valid gender_accuracy: 0.928 +-0.004 (17639) 2020-02-02 00:20:09.866: INFO @decay_lr : LR updated to `3.8389146e-05` 2020-02-02 00:20:10.199: INFO @save : Model saved to: ./log/GenderAgeClass_CleanAgeLabels_serene-bose/model_best_gender.ckpt 2020-02-02 00:20:10.203: INFO @log_profile : T train: 121.763343 2020-02-02 00:20:10.203: INFO @log_profile : T valid: 5.500405 2020-02-02 00:20:10.203: INFO @log_profile : T read data: 3.006221 2020-02-02 00:20:10.203: INFO @log_profile : T hooks: 5.465901 2020-02-02 00:20:10.203: INFO @main_loop : Epoch 191 done 2020-02-02 00:20:10.203: INFO @main_loop : Training epoch 192 2020-02-02 00:22:21.072: INFO @log_variables: train loss nanmean: 0.646097 2020-02-02 00:22:21.072: INFO @log_variables: train age_loss mean: 4.761034 2020-02-02 00:22:21.072: INFO @log_variables: train gender_loss mean: 0.100234 2020-02-02 00:22:21.072: INFO @log_variables: train age_mae mean: 5.234950 2020-02-02 00:22:21.072: INFO @log_variables: train gender_accuracy mean: 0.961313 2020-02-02 00:22:21.072: INFO @log_variables: train gender_confidence/loss nanmean: 0.050191 2020-02-02 00:22:21.072: INFO @log_variables: train gender_confidence/accuracy mean: 0.861686 2020-02-02 00:22:21.072: INFO @log_variables: train age_confidence/loss mean: 0.071962 2020-02-02 00:22:21.072: INFO @log_variables: train age_confidence/accuracy mean: 0.614114 2020-02-02 00:22:21.072: INFO @log_variables: valid loss nanmean: 0.853270 2020-02-02 00:22:21.072: INFO @log_variables: valid age_loss mean: 5.829268 2020-02-02 00:22:21.072: INFO @log_variables: valid gender_loss mean: 0.216680 2020-02-02 00:22:21.072: INFO @log_variables: valid age_mae mean: 6.308642 2020-02-02 00:22:21.072: INFO @log_variables: valid gender_accuracy mean: 0.921878 2020-02-02 00:22:21.072: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055912 2020-02-02 00:22:21.073: INFO @log_variables: valid gender_confidence/accuracy mean: 0.873349 2020-02-02 00:22:21.073: INFO @log_variables: valid age_confidence/loss mean: 0.070443 2020-02-02 00:22:21.073: INFO @log_variables: valid age_confidence/accuracy mean: 0.561880 2020-02-02 00:22:21.073: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:22:21.080: INFO @metrics_hook: train age_mae: 5.235 +-0.030 (110372) 2020-02-02 00:22:21.088: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:22:23.832: INFO @metrics_hook: valid age_mae: 6.309 +-0.090 (17639) 2020-02-02 00:22:23.834: INFO @metrics_hook: valid gender_accuracy: 0.922 +-0.004 (17639) 2020-02-02 00:22:25.317: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:22:25.317: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.83 +- 0.24 2020-02-02 00:22:25.317: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:22:25.317: INFO @evaluate_confidence: Average confidence of all samples 0.81 +- 0.26 2020-02-02 00:22:25.444: INFO @evaluate_confidence: Previous accuracy would be: 96.13 2020-02-02 00:22:25.444: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82] 2020-02-02 00:22:25.507: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.41, 98.48, 98.53, 98.58, 98.64, 98.69, 98.72, 98.77, 98.81, 98.84, 98.88, 98.91, 98.95, 98.98, 99.02, 99.07, 99.11, 99.14, 99.17, 99.19, 99.23, 99.24, 99.26, 99.29, 99.31, 99.33, 99.35, 99.37, 99.39, 99.42, 99.44, 99.46, 99.48, 99.49, 99.51, 99.53, 99.55, 99.57, 99.58, 99.6, 99.61, 99.63, 99.63, 99.66, 99.66, 99.68, 99.69, 99.7, 99.72, 99.74, 99.74] 2020-02-02 00:22:25.507: INFO @evaluate_confidence: Dropped ratios are: [9.53, 9.96, 10.35, 10.72, 11.11, 11.5, 11.85, 12.22, 12.58, 12.95, 13.35, 13.71, 14.06, 14.4, 14.81, 15.17, 15.56, 15.93, 16.32, 16.68, 17.05, 17.46, 17.82, 18.22, 18.59, 19.0, 19.39, 19.81, 20.2, 20.65, 21.05, 21.45, 21.91, 22.35, 22.82, 23.27, 23.77, 24.22, 24.71, 25.19, 25.72, 26.23, 26.77, 27.36, 27.94, 28.52, 29.14, 29.8, 30.46, 31.22, 31.96] 2020-02-02 00:22:25.557: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:22:25.557: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:22:25.558: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:22:25.558: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-02 00:22:25.699: INFO @evaluate_confidence: Previous accuracy would be: 60.58 2020-02-02 00:22:25.700: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:22:25.715: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.73, 69.39, 70.15, 70.89, 71.65, 72.48, 73.32] 2020-02-02 00:22:25.715: INFO @evaluate_confidence: Dropped ratios are: [42.48, 45.62, 48.66, 51.68, 54.56, 57.46, 60.2] 2020-02-02 00:22:25.723: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:22:25.723: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-02 00:22:25.723: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.46 +- 0.25 2020-02-02 00:22:25.724: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-02 00:22:25.834: INFO @evaluate_confidence: Previous accuracy would be: 92.19 2020-02-02 00:22:25.834: INFO @evaluate_confidence: Possible optimal thresholds are: [0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:22:25.843: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.31, 96.36, 96.44, 96.55, 96.63, 96.71, 96.8, 96.87, 96.95, 97.0, 97.05, 97.14, 97.22, 97.29, 97.35, 97.39, 97.46, 97.51, 97.58, 97.62, 97.67, 97.73, 97.76, 97.82, 97.91, 97.92, 98.02, 98.07, 98.12, 98.15, 98.25, 98.31, 98.35, 98.42, 98.48, 98.51, 98.54, 98.58, 98.65, 98.67, 98.71] 2020-02-02 00:22:25.844: INFO @evaluate_confidence: Dropped ratios are: [13.28, 13.61, 13.94, 14.33, 14.73, 15.11, 15.4, 15.74, 16.13, 16.51, 16.85, 17.23, 17.64, 18.01, 18.37, 18.68, 19.02, 19.43, 19.84, 20.19, 20.59, 20.98, 21.44, 21.89, 22.32, 22.68, 23.16, 23.58, 24.12, 24.63, 25.23, 25.72, 26.25, 26.82, 27.38, 28.02, 28.58, 29.17, 30.04, 30.76, 31.62] 2020-02-02 00:22:25.851: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:22:25.851: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.11 2020-02-02 00:22:25.851: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.10 2020-02-02 00:22:25.852: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.11 2020-02-02 00:22:25.980: INFO @evaluate_confidence: Previous accuracy would be: 52.49 2020-02-02 00:22:25.980: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-02 00:22:25.982: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.61, 58.08, 58.72, 59.2, 59.66] 2020-02-02 00:22:25.982: INFO @evaluate_confidence: Dropped ratios are: [45.95, 50.93, 55.75, 60.47, 64.95] 2020-02-02 00:22:26.039: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:22:26.784: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:22:26.880: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:22:27.382: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:22:27.651: INFO @decay_lr : LR updated to `3.81972e-05` 2020-02-02 00:22:27.652: INFO @log_profile : T train: 121.532691 2020-02-02 00:22:27.652: INFO @log_profile : T valid: 5.531429 2020-02-02 00:22:27.652: INFO @log_profile : T read data: 3.103047 2020-02-02 00:22:27.653: INFO @log_profile : T hooks: 7.203496 2020-02-02 00:22:27.653: INFO @main_loop : Epoch 192 done 2020-02-02 00:22:27.653: INFO @main_loop : Training epoch 193 2020-02-02 00:24:37.796: INFO @log_variables: train loss nanmean: 0.640728 2020-02-02 00:24:37.796: INFO @log_variables: train age_loss mean: 4.735009 2020-02-02 00:24:37.796: INFO @log_variables: train gender_loss mean: 0.098271 2020-02-02 00:24:37.796: INFO @log_variables: train age_mae mean: 5.208851 2020-02-02 00:24:37.796: INFO @log_variables: train gender_accuracy mean: 0.961941 2020-02-02 00:24:37.796: INFO @log_variables: train gender_confidence/loss nanmean: 0.048795 2020-02-02 00:24:37.796: INFO @log_variables: train gender_confidence/accuracy mean: 0.866663 2020-02-02 00:24:37.796: INFO @log_variables: train age_confidence/loss mean: 0.072141 2020-02-02 00:24:37.796: INFO @log_variables: train age_confidence/accuracy mean: 0.612784 2020-02-02 00:24:37.796: INFO @log_variables: valid loss nanmean: 0.840858 2020-02-02 00:24:37.796: INFO @log_variables: valid age_loss mean: 5.742784 2020-02-02 00:24:37.797: INFO @log_variables: valid gender_loss mean: 0.213944 2020-02-02 00:24:37.797: INFO @log_variables: valid age_mae mean: 6.223198 2020-02-02 00:24:37.797: INFO @log_variables: valid gender_accuracy mean: 0.924372 2020-02-02 00:24:37.797: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053906 2020-02-02 00:24:37.797: INFO @log_variables: valid gender_confidence/accuracy mean: 0.875050 2020-02-02 00:24:37.797: INFO @log_variables: valid age_confidence/loss mean: 0.070387 2020-02-02 00:24:37.797: INFO @log_variables: valid age_confidence/accuracy mean: 0.560859 2020-02-02 00:24:37.797: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:24:37.805: INFO @metrics_hook: train age_mae: 5.209 +-0.030 (110592) 2020-02-02 00:24:37.812: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110592) 2020-02-02 00:24:40.664: INFO @metrics_hook: valid age_mae: 6.223 +-0.088 (17639) 2020-02-02 00:24:40.665: INFO @metrics_hook: valid gender_accuracy: 0.924 +-0.004 (17639) 2020-02-02 00:24:42.420: INFO @decay_lr : LR updated to `3.8006212e-05` 2020-02-02 00:24:42.422: INFO @log_profile : T train: 121.970726 2020-02-02 00:24:42.422: INFO @log_profile : T valid: 5.514122 2020-02-02 00:24:42.422: INFO @log_profile : T read data: 1.950351 2020-02-02 00:24:42.422: INFO @log_profile : T hooks: 5.256318 2020-02-02 00:24:42.422: INFO @main_loop : Epoch 193 done 2020-02-02 00:24:42.422: INFO @main_loop : Training epoch 194 2020-02-02 00:26:53.208: INFO @log_variables: train loss nanmean: 0.641818 2020-02-02 00:26:53.208: INFO @log_variables: train age_loss mean: 4.754888 2020-02-02 00:26:53.208: INFO @log_variables: train gender_loss mean: 0.097343 2020-02-02 00:26:53.209: INFO @log_variables: train age_mae mean: 5.229398 2020-02-02 00:26:53.209: INFO @log_variables: train gender_accuracy mean: 0.962509 2020-02-02 00:26:53.209: INFO @log_variables: train gender_confidence/loss nanmean: 0.048990 2020-02-02 00:26:53.209: INFO @log_variables: train gender_confidence/accuracy mean: 0.865328 2020-02-02 00:26:53.209: INFO @log_variables: train age_confidence/loss mean: 0.072071 2020-02-02 00:26:53.209: INFO @log_variables: train age_confidence/accuracy mean: 0.613706 2020-02-02 00:26:53.209: INFO @log_variables: valid loss nanmean: 0.845414 2020-02-02 00:26:53.209: INFO @log_variables: valid age_loss mean: 5.786623 2020-02-02 00:26:53.209: INFO @log_variables: valid gender_loss mean: 0.213572 2020-02-02 00:26:53.209: INFO @log_variables: valid age_mae mean: 6.267019 2020-02-02 00:26:53.209: INFO @log_variables: valid gender_accuracy mean: 0.925223 2020-02-02 00:26:53.209: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055129 2020-02-02 00:26:53.209: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872158 2020-02-02 00:26:53.209: INFO @log_variables: valid age_confidence/loss mean: 0.070072 2020-02-02 00:26:53.209: INFO @log_variables: valid age_confidence/accuracy mean: 0.561880 2020-02-02 00:26:53.209: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:26:53.216: INFO @metrics_hook: train age_mae: 5.229 +-0.030 (110372) 2020-02-02 00:26:53.223: INFO @metrics_hook: train gender_accuracy: 0.963 +-0.001 (110372) 2020-02-02 00:26:55.915: INFO @metrics_hook: valid age_mae: 6.267 +-0.089 (17639) 2020-02-02 00:26:55.916: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-02 00:26:57.385: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:26:57.385: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:26:57.385: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:26:57.386: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:26:57.513: INFO @evaluate_confidence: Previous accuracy would be: 96.25 2020-02-02 00:26:57.513: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:26:57.577: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.43, 98.49, 98.55, 98.61, 98.66, 98.7, 98.75, 98.8, 98.84, 98.88, 98.93, 98.97, 99.0, 99.02, 99.06, 99.09, 99.12, 99.15, 99.19, 99.21, 99.24, 99.26, 99.28, 99.31, 99.34, 99.37, 99.39, 99.41, 99.43, 99.44, 99.47, 99.48, 99.5, 99.52, 99.54, 99.56, 99.58, 99.59, 99.59, 99.6, 99.61, 99.62, 99.65, 99.67, 99.69, 99.7, 99.71, 99.72, 99.73, 99.75, 99.78, 99.79] 2020-02-02 00:26:57.577: INFO @evaluate_confidence: Dropped ratios are: [9.17, 9.55, 9.92, 10.32, 10.71, 11.1, 11.47, 11.83, 12.2, 12.57, 12.94, 13.3, 13.66, 14.02, 14.4, 14.77, 15.11, 15.46, 15.85, 16.21, 16.57, 16.96, 17.32, 17.69, 18.08, 18.5, 18.88, 19.26, 19.66, 20.08, 20.47, 20.84, 21.28, 21.73, 22.19, 22.62, 23.08, 23.53, 23.98, 24.45, 24.93, 25.44, 26.0, 26.55, 27.13, 27.76, 28.4, 29.09, 29.75, 30.42, 31.23, 32.03] 2020-02-02 00:26:57.627: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:26:57.628: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:26:57.628: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:26:57.628: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-02 00:26:57.767: INFO @evaluate_confidence: Previous accuracy would be: 60.72 2020-02-02 00:26:57.768: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:26:57.782: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.63, 69.32, 70.05, 70.77, 71.6, 72.34, 73.21] 2020-02-02 00:26:57.782: INFO @evaluate_confidence: Dropped ratios are: [41.93, 45.18, 48.41, 51.44, 54.49, 57.35, 60.1] 2020-02-02 00:26:57.789: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:26:57.790: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.22 2020-02-02 00:26:57.790: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.45 +- 0.25 2020-02-02 00:26:57.790: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-02 00:26:57.898: INFO @evaluate_confidence: Previous accuracy would be: 92.52 2020-02-02 00:26:57.899: INFO @evaluate_confidence: Possible optimal thresholds are: [0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:26:57.908: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.51, 96.59, 96.66, 96.77, 96.86, 96.94, 97.02, 97.12, 97.17, 97.22, 97.27, 97.32, 97.35, 97.46, 97.53, 97.6, 97.67, 97.71, 97.83, 97.88, 97.9, 97.94, 98.0, 98.04, 98.07, 98.11, 98.18, 98.2, 98.22, 98.24, 98.3, 98.34, 98.4, 98.46, 98.49, 98.54, 98.62, 98.71, 98.76, 98.79, 98.84, 98.88] 2020-02-02 00:26:57.908: INFO @evaluate_confidence: Dropped ratios are: [13.09, 13.5, 13.86, 14.31, 14.68, 15.07, 15.44, 15.83, 16.19, 16.54, 16.85, 17.23, 17.52, 18.04, 18.36, 18.71, 19.19, 19.49, 19.87, 20.3, 20.74, 21.19, 21.61, 21.98, 22.43, 22.85, 23.33, 23.67, 24.08, 24.51, 24.95, 25.45, 25.99, 26.57, 27.18, 27.72, 28.42, 29.0, 29.59, 30.43, 31.18, 31.88] 2020-02-02 00:26:57.916: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:26:57.916: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.53 +- 0.11 2020-02-02 00:26:57.916: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.49 +- 0.09 2020-02-02 00:26:57.916: INFO @evaluate_confidence: Average confidence of all samples 0.51 +- 0.10 2020-02-02 00:26:58.042: INFO @evaluate_confidence: Previous accuracy would be: 52.48 2020-02-02 00:26:58.042: INFO @evaluate_confidence: Possible optimal thresholds are: [0.49, 0.5, 0.51, 0.52] 2020-02-02 00:26:58.044: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [58.13, 58.77, 59.85, 60.53] 2020-02-02 00:26:58.044: INFO @evaluate_confidence: Dropped ratios are: [45.44, 50.57, 56.01, 60.83] 2020-02-02 00:26:58.097: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:26:58.784: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:26:58.875: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:26:59.351: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:26:59.613: INFO @decay_lr : LR updated to `3.7816182e-05` 2020-02-02 00:26:59.614: INFO @log_profile : T train: 121.645085 2020-02-02 00:26:59.614: INFO @log_profile : T valid: 5.455511 2020-02-02 00:26:59.614: INFO @log_profile : T read data: 2.984026 2020-02-02 00:26:59.614: INFO @log_profile : T hooks: 7.029678 2020-02-02 00:26:59.615: INFO @main_loop : Epoch 194 done 2020-02-02 00:26:59.615: INFO @main_loop : Training epoch 195 2020-02-02 00:29:10.382: INFO @log_variables: train loss nanmean: 0.639207 2020-02-02 00:29:10.382: INFO @log_variables: train age_loss mean: 4.734630 2020-02-02 00:29:10.382: INFO @log_variables: train gender_loss mean: 0.095930 2020-02-02 00:29:10.383: INFO @log_variables: train age_mae mean: 5.208622 2020-02-02 00:29:10.383: INFO @log_variables: train gender_accuracy mean: 0.963496 2020-02-02 00:29:10.383: INFO @log_variables: train gender_confidence/loss nanmean: 0.049366 2020-02-02 00:29:10.383: INFO @log_variables: train gender_confidence/accuracy mean: 0.867068 2020-02-02 00:29:10.383: INFO @log_variables: train age_confidence/loss mean: 0.072210 2020-02-02 00:29:10.383: INFO @log_variables: train age_confidence/accuracy mean: 0.612284 2020-02-02 00:29:10.383: INFO @log_variables: valid loss nanmean: 0.834896 2020-02-02 00:29:10.383: INFO @log_variables: valid age_loss mean: 5.718418 2020-02-02 00:29:10.383: INFO @log_variables: valid gender_loss mean: 0.209933 2020-02-02 00:29:10.383: INFO @log_variables: valid age_mae mean: 6.198171 2020-02-02 00:29:10.383: INFO @log_variables: valid gender_accuracy mean: 0.924996 2020-02-02 00:29:10.383: INFO @log_variables: valid gender_confidence/loss nanmean: 0.053718 2020-02-02 00:29:10.383: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872498 2020-02-02 00:29:10.383: INFO @log_variables: valid age_confidence/loss mean: 0.070473 2020-02-02 00:29:10.383: INFO @log_variables: valid age_confidence/accuracy mean: 0.558875 2020-02-02 00:29:10.383: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:29:10.391: INFO @metrics_hook: train age_mae: 5.209 +-0.030 (110372) 2020-02-02 00:29:10.398: INFO @metrics_hook: train gender_accuracy: 0.963 +-0.001 (110372) 2020-02-02 00:29:13.174: INFO @metrics_hook: valid age_mae: 6.198 +-0.087 (17639) 2020-02-02 00:29:13.176: INFO @metrics_hook: valid gender_accuracy: 0.925 +-0.004 (17639) 2020-02-02 00:29:14.865: INFO @decay_lr : LR updated to `3.7627102e-05` 2020-02-02 00:29:14.866: INFO @log_profile : T train: 121.728124 2020-02-02 00:29:14.866: INFO @log_profile : T valid: 5.488922 2020-02-02 00:29:14.866: INFO @log_profile : T read data: 2.856607 2020-02-02 00:29:14.866: INFO @log_profile : T hooks: 5.100759 2020-02-02 00:29:14.866: INFO @main_loop : Epoch 195 done 2020-02-02 00:29:14.867: INFO @main_loop : Training epoch 196 2020-02-02 00:31:24.744: INFO @log_variables: train loss nanmean: 0.633949 2020-02-02 00:31:24.745: INFO @log_variables: train age_loss mean: 4.698214 2020-02-02 00:31:24.745: INFO @log_variables: train gender_loss mean: 0.093955 2020-02-02 00:31:24.745: INFO @log_variables: train age_mae mean: 5.171704 2020-02-02 00:31:24.745: INFO @log_variables: train gender_accuracy mean: 0.963569 2020-02-02 00:31:24.745: INFO @log_variables: train gender_confidence/loss nanmean: 0.049202 2020-02-02 00:31:24.745: INFO @log_variables: train gender_confidence/accuracy mean: 0.867576 2020-02-02 00:31:24.745: INFO @log_variables: train age_confidence/loss mean: 0.072224 2020-02-02 00:31:24.745: INFO @log_variables: train age_confidence/accuracy mean: 0.613064 2020-02-02 00:31:24.745: INFO @log_variables: valid loss nanmean: 0.845300 2020-02-02 00:31:24.745: INFO @log_variables: valid age_loss mean: 5.798472 2020-02-02 00:31:24.745: INFO @log_variables: valid gender_loss mean: 0.211755 2020-02-02 00:31:24.745: INFO @log_variables: valid age_mae mean: 6.279001 2020-02-02 00:31:24.745: INFO @log_variables: valid gender_accuracy mean: 0.925960 2020-02-02 00:31:24.745: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055438 2020-02-02 00:31:24.745: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871138 2020-02-02 00:31:24.745: INFO @log_variables: valid age_confidence/loss mean: 0.070223 2020-02-02 00:31:24.745: INFO @log_variables: valid age_confidence/accuracy mean: 0.554680 2020-02-02 00:31:24.746: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:31:24.753: INFO @metrics_hook: train age_mae: 5.172 +-0.030 (110592) 2020-02-02 00:31:24.760: INFO @metrics_hook: train gender_accuracy: 0.964 +-0.001 (110592) 2020-02-02 00:31:27.488: INFO @metrics_hook: valid age_mae: 6.279 +-0.089 (17639) 2020-02-02 00:31:27.489: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-02 00:31:28.941: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:31:28.941: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:31:28.941: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.22 2020-02-02 00:31:28.941: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:31:29.068: INFO @evaluate_confidence: Previous accuracy would be: 96.36 2020-02-02 00:31:29.068: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:31:29.132: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.5, 98.55, 98.59, 98.64, 98.71, 98.75, 98.78, 98.83, 98.86, 98.9, 98.94, 98.98, 99.01, 99.03, 99.06, 99.09, 99.11, 99.15, 99.17, 99.2, 99.23, 99.26, 99.28, 99.31, 99.34, 99.36, 99.38, 99.4, 99.42, 99.43, 99.46, 99.49, 99.52, 99.53, 99.54, 99.56, 99.57, 99.58, 99.61, 99.62, 99.64, 99.65, 99.66, 99.67, 99.68, 99.7, 99.71, 99.73, 99.74, 99.76, 99.77, 99.78] 2020-02-02 00:31:29.132: INFO @evaluate_confidence: Dropped ratios are: [9.04, 9.39, 9.76, 10.15, 10.56, 10.9, 11.25, 11.59, 11.95, 12.29, 12.64, 12.99, 13.37, 13.72, 14.03, 14.39, 14.74, 15.1, 15.48, 15.82, 16.19, 16.52, 16.9, 17.23, 17.61, 17.98, 18.39, 18.75, 19.11, 19.5, 19.92, 20.34, 20.75, 21.18, 21.58, 22.05, 22.51, 22.95, 23.45, 23.93, 24.42, 24.94, 25.48, 26.01, 26.59, 27.2, 27.78, 28.41, 29.06, 29.79, 30.52, 31.29] 2020-02-02 00:31:29.180: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:31:29.181: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:31:29.181: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:31:29.181: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:31:29.323: INFO @evaluate_confidence: Previous accuracy would be: 61.03 2020-02-02 00:31:29.323: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:31:29.338: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.78, 69.46, 70.18, 70.96, 71.66, 72.53, 73.36] 2020-02-02 00:31:29.338: INFO @evaluate_confidence: Dropped ratios are: [41.77, 44.93, 48.16, 51.24, 54.22, 57.02, 59.74] 2020-02-02 00:31:29.346: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:31:29.346: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.23 2020-02-02 00:31:29.346: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.44 +- 0.26 2020-02-02 00:31:29.346: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.26 2020-02-02 00:31:29.452: INFO @evaluate_confidence: Previous accuracy would be: 92.60 2020-02-02 00:31:29.452: INFO @evaluate_confidence: Possible optimal thresholds are: [0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:31:29.462: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.6, 96.65, 96.78, 96.84, 96.91, 96.97, 97.04, 97.12, 97.22, 97.32, 97.39, 97.48, 97.53, 97.61, 97.68, 97.72, 97.74, 97.76, 97.81, 97.85, 97.91, 97.96, 98.0, 98.04, 98.06, 98.11, 98.14, 98.17, 98.19, 98.25, 98.27, 98.3, 98.33, 98.37, 98.42, 98.48, 98.48, 98.55, 98.6, 98.61, 98.69, 98.72, 98.8] 2020-02-02 00:31:29.462: INFO @evaluate_confidence: Dropped ratios are: [13.24, 13.56, 13.97, 14.31, 14.63, 14.93, 15.27, 15.66, 16.08, 16.49, 16.93, 17.29, 17.68, 18.07, 18.5, 18.87, 19.19, 19.41, 19.87, 20.15, 20.61, 20.99, 21.34, 21.72, 22.06, 22.47, 22.89, 23.31, 23.78, 24.38, 24.88, 25.44, 25.86, 26.45, 26.95, 27.54, 28.18, 28.82, 29.57, 30.22, 30.9, 31.56, 32.31] 2020-02-02 00:31:29.469: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:31:29.469: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.11 2020-02-02 00:31:29.470: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-02 00:31:29.470: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-02 00:31:29.597: INFO @evaluate_confidence: Previous accuracy would be: 52.53 2020-02-02 00:31:29.597: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-02 00:31:29.599: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.18, 58.04, 58.99, 60.15, 60.97] 2020-02-02 00:31:29.599: INFO @evaluate_confidence: Dropped ratios are: [44.4, 49.21, 54.21, 59.14, 64.02] 2020-02-02 00:31:29.651: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:31:30.342: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:31:30.428: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:31:30.907: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:31:31.170: INFO @decay_lr : LR updated to `3.7438967e-05` 2020-02-02 00:31:31.171: INFO @log_profile : T train: 121.732713 2020-02-02 00:31:31.171: INFO @log_profile : T valid: 5.453504 2020-02-02 00:31:31.171: INFO @log_profile : T read data: 1.986358 2020-02-02 00:31:31.171: INFO @log_profile : T hooks: 7.053678 2020-02-02 00:31:31.171: INFO @main_loop : Epoch 196 done 2020-02-02 00:31:31.171: INFO @main_loop : Training epoch 197 2020-02-02 00:33:49.236: INFO @log_variables: train loss nanmean: 0.639448 2020-02-02 00:33:49.236: INFO @log_variables: train age_loss mean: 4.708799 2020-02-02 00:33:49.236: INFO @log_variables: train gender_loss mean: 0.099106 2020-02-02 00:33:49.236: INFO @log_variables: train age_mae mean: 5.182750 2020-02-02 00:33:49.236: INFO @log_variables: train gender_accuracy mean: 0.961403 2020-02-02 00:33:49.236: INFO @log_variables: train gender_confidence/loss nanmean: 0.049020 2020-02-02 00:33:49.236: INFO @log_variables: train gender_confidence/accuracy mean: 0.865546 2020-02-02 00:33:49.236: INFO @log_variables: train age_confidence/loss mean: 0.072259 2020-02-02 00:33:49.236: INFO @log_variables: train age_confidence/accuracy mean: 0.616905 2020-02-02 00:33:49.236: INFO @log_variables: valid loss nanmean: 0.839170 2020-02-02 00:33:49.236: INFO @log_variables: valid age_loss mean: 5.755429 2020-02-02 00:33:49.236: INFO @log_variables: valid gender_loss mean: 0.210069 2020-02-02 00:33:49.236: INFO @log_variables: valid age_mae mean: 6.236133 2020-02-02 00:33:49.237: INFO @log_variables: valid gender_accuracy mean: 0.925846 2020-02-02 00:33:49.237: INFO @log_variables: valid gender_confidence/loss nanmean: 0.054649 2020-02-02 00:33:49.237: INFO @log_variables: valid gender_confidence/accuracy mean: 0.869891 2020-02-02 00:33:49.237: INFO @log_variables: valid age_confidence/loss mean: 0.070328 2020-02-02 00:33:49.237: INFO @log_variables: valid age_confidence/accuracy mean: 0.557458 2020-02-02 00:33:49.237: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:33:49.244: INFO @metrics_hook: train age_mae: 5.183 +-0.030 (110372) 2020-02-02 00:33:49.251: INFO @metrics_hook: train gender_accuracy: 0.961 +-0.001 (110372) 2020-02-02 00:33:51.972: INFO @metrics_hook: valid age_mae: 6.236 +-0.089 (17639) 2020-02-02 00:33:51.973: INFO @metrics_hook: valid gender_accuracy: 0.926 +-0.004 (17639) 2020-02-02 00:33:53.635: INFO @decay_lr : LR updated to `3.725177e-05` 2020-02-02 00:33:53.636: INFO @log_profile : T train: 129.001735 2020-02-02 00:33:53.636: INFO @log_profile : T valid: 5.603191 2020-02-02 00:33:53.636: INFO @log_profile : T read data: 2.787573 2020-02-02 00:33:53.636: INFO @log_profile : T hooks: 4.997001 2020-02-02 00:33:53.636: INFO @main_loop : Epoch 197 done 2020-02-02 00:33:53.636: INFO @main_loop : Training epoch 198 2020-02-02 00:36:05.621: INFO @log_variables: train loss nanmean: 0.636567 2020-02-02 00:36:05.621: INFO @log_variables: train age_loss mean: 4.718603 2020-02-02 00:36:05.621: INFO @log_variables: train gender_loss mean: 0.095499 2020-02-02 00:36:05.622: INFO @log_variables: train age_mae mean: 5.192708 2020-02-02 00:36:05.622: INFO @log_variables: train gender_accuracy mean: 0.962500 2020-02-02 00:36:05.622: INFO @log_variables: train gender_confidence/loss nanmean: 0.048467 2020-02-02 00:36:05.622: INFO @log_variables: train gender_confidence/accuracy mean: 0.868055 2020-02-02 00:36:05.622: INFO @log_variables: train age_confidence/loss mean: 0.072319 2020-02-02 00:36:05.622: INFO @log_variables: train age_confidence/accuracy mean: 0.616406 2020-02-02 00:36:05.622: INFO @log_variables: valid loss nanmean: 0.864424 2020-02-02 00:36:05.622: INFO @log_variables: valid age_loss mean: 5.783998 2020-02-02 00:36:05.622: INFO @log_variables: valid gender_loss mean: 0.232744 2020-02-02 00:36:05.622: INFO @log_variables: valid age_mae mean: 6.263634 2020-02-02 00:36:05.622: INFO @log_variables: valid gender_accuracy mean: 0.920857 2020-02-02 00:36:05.622: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056441 2020-02-02 00:36:05.622: INFO @log_variables: valid gender_confidence/accuracy mean: 0.866489 2020-02-02 00:36:05.622: INFO @log_variables: valid age_confidence/loss mean: 0.070580 2020-02-02 00:36:05.622: INFO @log_variables: valid age_confidence/accuracy mean: 0.561370 2020-02-02 00:36:05.622: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:36:05.629: INFO @metrics_hook: train age_mae: 5.193 +-0.030 (110372) 2020-02-02 00:36:05.636: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110372) 2020-02-02 00:36:08.437: INFO @metrics_hook: valid age_mae: 6.264 +-0.089 (17639) 2020-02-02 00:36:08.438: INFO @metrics_hook: valid gender_accuracy: 0.921 +-0.004 (17639) 2020-02-02 00:36:09.921: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:36:09.922: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:36:09.922: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:36:09.922: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:36:10.050: INFO @evaluate_confidence: Previous accuracy would be: 96.25 2020-02-02 00:36:10.051: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:36:10.114: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.45, 98.51, 98.55, 98.6, 98.66, 98.7, 98.75, 98.79, 98.83, 98.88, 98.91, 98.96, 98.99, 99.02, 99.06, 99.09, 99.12, 99.15, 99.18, 99.21, 99.23, 99.25, 99.27, 99.3, 99.33, 99.37, 99.39, 99.42, 99.44, 99.47, 99.48, 99.5, 99.52, 99.54, 99.55, 99.57, 99.59, 99.61, 99.62, 99.64, 99.65, 99.66, 99.67, 99.68, 99.7, 99.71, 99.72, 99.74, 99.75, 99.76, 99.77, 99.78] 2020-02-02 00:36:10.115: INFO @evaluate_confidence: Dropped ratios are: [9.05, 9.4, 9.77, 10.13, 10.52, 10.86, 11.21, 11.56, 11.94, 12.29, 12.64, 13.04, 13.4, 13.75, 14.13, 14.46, 14.83, 15.2, 15.56, 15.91, 16.29, 16.62, 16.98, 17.35, 17.72, 18.09, 18.5, 18.89, 19.32, 19.71, 20.12, 20.53, 20.94, 21.36, 21.8, 22.3, 22.78, 23.28, 23.78, 24.24, 24.74, 25.26, 25.77, 26.37, 26.97, 27.54, 28.17, 28.82, 29.44, 30.13, 30.85, 31.62] 2020-02-02 00:36:10.165: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:36:10.165: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:36:10.166: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:36:10.166: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.13 2020-02-02 00:36:10.306: INFO @evaluate_confidence: Previous accuracy would be: 61.12 2020-02-02 00:36:10.307: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:36:10.321: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [68.94, 69.59, 70.29, 71.08, 71.78, 72.58, 73.43] 2020-02-02 00:36:10.322: INFO @evaluate_confidence: Dropped ratios are: [40.99, 44.1, 47.2, 50.34, 53.4, 56.44, 59.22] 2020-02-02 00:36:10.329: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:36:10.329: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.23 2020-02-02 00:36:10.329: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.47 +- 0.27 2020-02-02 00:36:10.329: INFO @evaluate_confidence: Average confidence of all samples 0.83 +- 0.25 2020-02-02 00:36:10.433: INFO @evaluate_confidence: Previous accuracy would be: 92.09 2020-02-02 00:36:10.434: INFO @evaluate_confidence: Possible optimal thresholds are: [0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:36:10.442: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.18, 96.28, 96.37, 96.45, 96.53, 96.61, 96.65, 96.69, 96.74, 96.79, 96.94, 96.98, 97.04, 97.09, 97.14, 97.21, 97.24, 97.29, 97.35, 97.4, 97.44, 97.51, 97.57, 97.59, 97.62, 97.67, 97.73, 97.79, 97.86, 97.93, 98.0, 98.05, 98.1, 98.17, 98.2, 98.27, 98.31, 98.34, 98.37, 98.42] 2020-02-02 00:36:10.442: INFO @evaluate_confidence: Dropped ratios are: [14.02, 14.42, 14.83, 15.24, 15.64, 15.98, 16.33, 16.62, 16.92, 17.29, 17.67, 18.08, 18.46, 18.83, 19.24, 19.65, 19.96, 20.31, 20.66, 21.08, 21.55, 21.95, 22.34, 22.71, 23.17, 23.58, 24.09, 24.59, 25.11, 25.57, 26.07, 26.56, 27.09, 27.68, 28.16, 28.74, 29.4, 30.04, 30.67, 31.41] 2020-02-02 00:36:10.450: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:36:10.450: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.55 +- 0.10 2020-02-02 00:36:10.450: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.09 2020-02-02 00:36:10.450: INFO @evaluate_confidence: Average confidence of all samples 0.53 +- 0.10 2020-02-02 00:36:10.579: INFO @evaluate_confidence: Previous accuracy would be: 52.90 2020-02-02 00:36:10.579: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55] 2020-02-02 00:36:10.580: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.45, 57.84, 58.63, 59.66, 60.52] 2020-02-02 00:36:10.581: INFO @evaluate_confidence: Dropped ratios are: [43.53, 48.82, 53.87, 59.46, 64.29] 2020-02-02 00:36:10.635: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:36:11.338: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:36:11.425: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:36:11.897: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:36:12.167: INFO @decay_lr : LR updated to `3.7065514e-05` 2020-02-02 00:36:12.169: INFO @log_profile : T train: 122.871105 2020-02-02 00:36:12.169: INFO @log_profile : T valid: 5.559683 2020-02-02 00:36:12.169: INFO @log_profile : T read data: 2.866097 2020-02-02 00:36:12.169: INFO @log_profile : T hooks: 7.159736 2020-02-02 00:36:12.169: INFO @main_loop : Epoch 198 done 2020-02-02 00:36:12.169: INFO @main_loop : Training epoch 199 2020-02-02 00:38:31.335: INFO @log_variables: train loss nanmean: 0.639718 2020-02-02 00:38:31.336: INFO @log_variables: train age_loss mean: 4.742259 2020-02-02 00:38:31.336: INFO @log_variables: train gender_loss mean: 0.096016 2020-02-02 00:38:31.336: INFO @log_variables: train age_mae mean: 5.216309 2020-02-02 00:38:31.336: INFO @log_variables: train gender_accuracy mean: 0.962864 2020-02-02 00:38:31.336: INFO @log_variables: train gender_confidence/loss nanmean: 0.049203 2020-02-02 00:38:31.336: INFO @log_variables: train gender_confidence/accuracy mean: 0.866916 2020-02-02 00:38:31.336: INFO @log_variables: train age_confidence/loss mean: 0.072113 2020-02-02 00:38:31.336: INFO @log_variables: train age_confidence/accuracy mean: 0.616148 2020-02-02 00:38:31.336: INFO @log_variables: valid loss nanmean: 0.843325 2020-02-02 00:38:31.336: INFO @log_variables: valid age_loss mean: 5.762456 2020-02-02 00:38:31.336: INFO @log_variables: valid gender_loss mean: 0.212079 2020-02-02 00:38:31.336: INFO @log_variables: valid age_mae mean: 6.243574 2020-02-02 00:38:31.336: INFO @log_variables: valid gender_accuracy mean: 0.928227 2020-02-02 00:38:31.336: INFO @log_variables: valid gender_confidence/loss nanmean: 0.056486 2020-02-02 00:38:31.336: INFO @log_variables: valid gender_confidence/accuracy mean: 0.871308 2020-02-02 00:38:31.336: INFO @log_variables: valid age_confidence/loss mean: 0.070180 2020-02-02 00:38:31.336: INFO @log_variables: valid age_confidence/accuracy mean: 0.555814 2020-02-02 00:38:31.336: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:38:31.343: INFO @metrics_hook: train age_mae: 5.216 +-0.030 (110592) 2020-02-02 00:38:31.350: INFO @metrics_hook: train gender_accuracy: 0.963 +-0.001 (110592) 2020-02-02 00:38:34.054: INFO @metrics_hook: valid age_mae: 6.244 +-0.089 (17639) 2020-02-02 00:38:34.055: INFO @metrics_hook: valid gender_accuracy: 0.928 +-0.004 (17639) 2020-02-02 00:38:35.707: INFO @decay_lr : LR updated to `3.688019e-05` 2020-02-02 00:38:35.708: INFO @log_profile : T train: 130.146162 2020-02-02 00:38:35.709: INFO @log_profile : T valid: 6.411105 2020-02-02 00:38:35.709: INFO @log_profile : T read data: 1.917038 2020-02-02 00:38:35.709: INFO @log_profile : T hooks: 4.989021 2020-02-02 00:38:35.709: INFO @main_loop : Epoch 199 done 2020-02-02 00:38:35.709: INFO @main_loop : Training epoch 200 2020-02-02 00:40:55.517: INFO @log_variables: train loss nanmean: 0.637628 2020-02-02 00:40:55.517: INFO @log_variables: train age_loss mean: 4.715021 2020-02-02 00:40:55.517: INFO @log_variables: train gender_loss mean: 0.097335 2020-02-02 00:40:55.517: INFO @log_variables: train age_mae mean: 5.189459 2020-02-02 00:40:55.517: INFO @log_variables: train gender_accuracy mean: 0.962427 2020-02-02 00:40:55.517: INFO @log_variables: train gender_confidence/loss nanmean: 0.048258 2020-02-02 00:40:55.517: INFO @log_variables: train gender_confidence/accuracy mean: 0.864766 2020-02-02 00:40:55.517: INFO @log_variables: train age_confidence/loss mean: 0.072245 2020-02-02 00:40:55.517: INFO @log_variables: train age_confidence/accuracy mean: 0.615691 2020-02-02 00:40:55.517: INFO @log_variables: valid loss nanmean: 0.838027 2020-02-02 00:40:55.517: INFO @log_variables: valid age_loss mean: 5.788793 2020-02-02 00:40:55.517: INFO @log_variables: valid gender_loss mean: 0.205341 2020-02-02 00:40:55.517: INFO @log_variables: valid age_mae mean: 6.269044 2020-02-02 00:40:55.517: INFO @log_variables: valid gender_accuracy mean: 0.926526 2020-02-02 00:40:55.517: INFO @log_variables: valid gender_confidence/loss nanmean: 0.055063 2020-02-02 00:40:55.517: INFO @log_variables: valid gender_confidence/accuracy mean: 0.872045 2020-02-02 00:40:55.517: INFO @log_variables: valid age_confidence/loss mean: 0.070036 2020-02-02 00:40:55.518: INFO @log_variables: valid age_confidence/accuracy mean: 0.560689 2020-02-02 00:40:55.518: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose 2020-02-02 00:40:55.525: INFO @metrics_hook: train age_mae: 5.189 +-0.030 (110372) 2020-02-02 00:40:55.532: INFO @metrics_hook: train gender_accuracy: 0.962 +-0.001 (110372) 2020-02-02 00:40:58.270: INFO @metrics_hook: valid age_mae: 6.269 +-0.089 (17639) 2020-02-02 00:40:58.271: INFO @metrics_hook: valid gender_accuracy: 0.927 +-0.004 (17639) 2020-02-02 00:40:59.721: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:40:59.722: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.84 +- 0.24 2020-02-02 00:40:59.722: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.32 +- 0.21 2020-02-02 00:40:59.722: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.26 2020-02-02 00:40:59.849: INFO @evaluate_confidence: Previous accuracy would be: 96.24 2020-02-02 00:40:59.850: INFO @evaluate_confidence: Possible optimal thresholds are: [0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4, 0.41, 0.42, 0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83] 2020-02-02 00:40:59.914: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [98.47, 98.52, 98.59, 98.63, 98.67, 98.72, 98.77, 98.81, 98.86, 98.9, 98.94, 98.98, 99.02, 99.06, 99.09, 99.12, 99.15, 99.17, 99.2, 99.22, 99.24, 99.27, 99.29, 99.31, 99.33, 99.36, 99.39, 99.41, 99.43, 99.45, 99.48, 99.49, 99.51, 99.53, 99.56, 99.58, 99.59, 99.61, 99.64, 99.65, 99.66, 99.68, 99.69, 99.71, 99.72, 99.73, 99.74, 99.76, 99.77, 99.78, 99.79, 99.8] 2020-02-02 00:40:59.914: INFO @evaluate_confidence: Dropped ratios are: [9.4, 9.8, 10.15, 10.52, 10.89, 11.25, 11.62, 11.96, 12.32, 12.7, 13.03, 13.4, 13.78, 14.15, 14.51, 14.85, 15.21, 15.56, 15.94, 16.31, 16.67, 16.99, 17.34, 17.73, 18.09, 18.48, 18.9, 19.25, 19.64, 20.05, 20.46, 20.85, 21.26, 21.7, 22.12, 22.55, 23.05, 23.51, 23.98, 24.5, 25.0, 25.52, 26.07, 26.58, 27.14, 27.74, 28.35, 29.03, 29.69, 30.38, 31.08, 31.81] 2020-02-02 00:40:59.964: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:40:59.964: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.58 +- 0.14 2020-02-02 00:40:59.964: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.51 +- 0.12 2020-02-02 00:40:59.965: INFO @evaluate_confidence: Average confidence of all samples 0.55 +- 0.14 2020-02-02 00:41:00.105: INFO @evaluate_confidence: Previous accuracy would be: 61.05 2020-02-02 00:41:00.105: INFO @evaluate_confidence: Possible optimal thresholds are: [0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57] 2020-02-02 00:41:00.120: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [69.1, 69.7, 70.38, 71.11, 71.88, 72.6, 73.35] 2020-02-02 00:41:00.121: INFO @evaluate_confidence: Dropped ratios are: [41.76, 44.85, 47.89, 50.89, 53.77, 56.58, 59.21] 2020-02-02 00:41:00.128: INFO @evaluate_confidence: Evaluating GENDER Confidence 2020-02-02 00:41:00.128: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.86 +- 0.23 2020-02-02 00:41:00.128: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.43 +- 0.26 2020-02-02 00:41:00.128: INFO @evaluate_confidence: Average confidence of all samples 0.82 +- 0.25 2020-02-02 00:41:00.235: INFO @evaluate_confidence: Previous accuracy would be: 92.65 2020-02-02 00:41:00.236: INFO @evaluate_confidence: Possible optimal thresholds are: [0.43, 0.44, 0.45, 0.46, 0.47, 0.48, 0.49, 0.5, 0.51, 0.52, 0.53, 0.54, 0.55, 0.56, 0.57, 0.58, 0.59, 0.6, 0.61, 0.62, 0.63, 0.64, 0.65, 0.66, 0.67, 0.68, 0.69, 0.7, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86] 2020-02-02 00:41:00.245: INFO @evaluate_confidence: GENDER accuracies with the thresholds are: [96.61, 96.73, 96.84, 96.89, 96.99, 97.08, 97.16, 97.22, 97.27, 97.36, 97.44, 97.49, 97.54, 97.59, 97.62, 97.66, 97.7, 97.77, 97.81, 97.83, 97.86, 97.93, 97.97, 98.0, 98.06, 98.1, 98.16, 98.2, 98.24, 98.28, 98.31, 98.33, 98.4, 98.44, 98.47, 98.5, 98.54, 98.57, 98.64, 98.7, 98.74, 98.75, 98.77, 98.78] 2020-02-02 00:41:00.245: INFO @evaluate_confidence: Dropped ratios are: [12.97, 13.31, 13.71, 14.04, 14.38, 14.73, 15.07, 15.45, 15.8, 16.16, 16.5, 16.89, 17.35, 17.77, 18.09, 18.5, 18.84, 19.25, 19.62, 19.93, 20.36, 20.82, 21.16, 21.57, 21.96, 22.38, 22.88, 23.3, 23.76, 24.19, 24.66, 25.11, 25.59, 26.15, 26.67, 27.27, 27.8, 28.43, 29.17, 29.8, 30.46, 31.16, 31.97, 32.75] 2020-02-02 00:41:00.252: INFO @evaluate_confidence: Evaluating AGE Confidence 2020-02-02 00:41:00.253: INFO @evaluate_confidence: Average confidence of correctly predicted samples 0.54 +- 0.10 2020-02-02 00:41:00.253: INFO @evaluate_confidence: Average confidence of wrongly predicted samples 0.50 +- 0.09 2020-02-02 00:41:00.253: INFO @evaluate_confidence: Average confidence of all samples 0.52 +- 0.10 2020-02-02 00:41:00.377: INFO @evaluate_confidence: Previous accuracy would be: 52.36 2020-02-02 00:41:00.378: INFO @evaluate_confidence: Possible optimal thresholds are: [0.5, 0.51, 0.52, 0.53, 0.54] 2020-02-02 00:41:00.379: INFO @evaluate_confidence: AGE accuracies with the thresholds are: [57.44, 57.9, 58.27, 58.9, 59.81] 2020-02-02 00:41:00.379: INFO @evaluate_confidence: Dropped ratios are: [43.42, 48.32, 53.31, 58.23, 63.21] 2020-02-02 00:41:00.431: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:41:01.123: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:41:01.210: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:41:01.672: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:41:01.745: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream train 2020-02-02 00:41:02.431: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:41:02.516: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-02 00:41:02.518: INFO @evaluate_gender-age_model: groups 0 3.153980 1 3.506725 2 4.848682 3 5.333799 4 5.938581 5 5.850243 6 5.789649 7 6.541146 Name: errors, dtype: float64 2020-02-02 00:41:02.519: INFO @evaluate_gender-age_model: Evaluating gender-age errors for stream valid 2020-02-02 00:41:02.984: INFO @evaluate_gender-age_model: Calculating age mean errors grouped by age ranges of 10 2020-02-02 00:41:03.048: INFO @evaluate_gender-age_model: Errors by groups: 2020-02-02 00:41:03.049: INFO @evaluate_gender-age_model: groups 0 5.584594 1 5.316784 2 5.757775 3 5.697492 4 7.231584 5 5.464110 6 7.674852 7 11.397805 Name: errors, dtype: float64 2020-02-02 00:41:03.240: INFO @decay_lr : LR updated to `3.6695787e-05` 2020-02-02 00:41:03.240: INFO @stop_after : EpochStopperHook triggered 2020-02-02 00:41:03.241: INFO @log_profile : T train: 128.861960 2020-02-02 00:41:03.241: INFO @log_profile : T valid: 5.595885 2020-02-02 00:41:03.241: INFO @log_profile : T read data: 2.803730 2020-02-02 00:41:03.241: INFO @log_profile : T hooks: 10.195249 2020-02-02 00:41:03.241: INFO @main_loop : Training terminated: Training terminated after epoch 200 2020-02-02 00:41:03.242: INFO @log_dir : Output dir: ./log/GenderAgeClass_CleanAgeLabels_serene-bose