Refactoring of experiment_logger into a class.

core/logger.py

@@ -3,15 +3,19 @@ import os
 import pickle
 
 
-def experiment_logger(rootdir, logdir_prefix, exp_config, model, training_df, test_df):
-    out_dir = os.path.join(rootdir, logdir_prefix)
+class Logger:
+    def __init__(self, rootdir):
+        self.rootdir = rootdir
 
-    # save trained model, exp_config.json, training_df, test_df, and computed errors on test set
-    os.makedirs(name=out_dir, exist_ok=True)
-    model.save(filepath=os.path.join(out_dir, 'model'))
-    with open(os.path.join(out_dir, 'exp_config.json'), mode='wt') as exp_config_f:
-        json.dump(exp_config, fp=exp_config_f, indent=4, sort_keys=True)
-    with open(os.path.join(out_dir, 'training_df.p'), mode='wb') as training_df_f:
-        pickle.dump(training_df, file=training_df_f)
-    with open(os.path.join(out_dir, 'test_df.p'), mode='wb') as test_df_f:
-        pickle.dump(test_df, file=test_df_f)
+    def log_model_training(self, logdir_prefix, exp_config, model, training_df, test_df):
+        out_dir = os.path.join(self.rootdir, logdir_prefix)
+
+        # save trained model, exp_config.json, training_df, test_df, and computed errors on test set
+        os.makedirs(name=out_dir, exist_ok=True)
+        model.save(filepath=os.path.join(out_dir, 'model'))
+        with open(os.path.join(out_dir, 'exp_config.json'), mode='wt') as exp_config_f:
+            json.dump(exp_config, fp=exp_config_f, indent=4, sort_keys=True)
+        with open(os.path.join(out_dir, 'training_df.p'), mode='wb') as training_df_f:
+            pickle.dump(training_df, file=training_df_f)
+        with open(os.path.join(out_dir, 'test_df.p'), mode='wb') as test_df_f:
+            pickle.dump(test_df, file=test_df_f)

scripts/train_model.py

@@ -4,7 +4,7 @@ from core.utils import load_ircam_hrirs_data, split_dataset, generate_signals, g
     azimuth_to_left_center_right_onehot
 import numpy as np
 import tensorflow as tf
-from core.logger import experiment_logger
+from core.logger import Logger
 from core.model import model_factory
 import pandas as pd
 from core.metrics import MeanAbsoluteAzimuthError
@@ -21,6 +21,7 @@ ExperimentConfiguration = dict
 
 ALL_SUBJECTS_EXCEPT_1059 = list(range(1002, 1010)) + list(range(1012, 1019)) + list(range(1021, 1024)) + list(
     [1025, 1026]) + list(range(1028, 1035)) + list(range(1037, 1059))
+RESULTS_ROOTDIR = '../tf_results'
 
 exp_config = ExperimentConfiguration({'n_augment': 1,
                                       'n_epochs': 20,
@@ -141,6 +142,7 @@ if __name__ == '__main__':
         timestr = datetime.today().strftime("%Y-%m-%d_%H-%M-%S")
         hrir_status = 'comp' if exp_config['compensated_hrirs'] else 'raw'
         logdir_prefix = f"{hrir_status}_{timestr}_{k}"
+        logger = Logger(rootdir=RESULTS_ROOTDIR)
 
         # create and train the model
         model = model_factory(model_name=exp_config['model_name'], model_config=exp_config['model_config'])
@@ -197,10 +199,19 @@ if __name__ == '__main__':
             # Here, width=time axis, and Height=channel (left or right)
             training_signals = np.expand_dims(training_signals, axis=-1)
             test_signals = np.expand_dims(test_signals, axis=-1)
-            model.fit(x=training_signals, y=training_labels, batch_size=batch_size, epochs=exp_config['n_epochs'],
-                      validation_data=(test_signals, test_labels))
+            model.fit(
+                x=training_signals, y=training_labels,
+                batch_size=batch_size,
+                epochs=exp_config['n_epochs'],
+                validation_data=(test_signals, test_labels),
+                callbacks=[tensorboard_callback],
+            )
 
         # log the experiment configuration file, the trained model, the training set and the test set
-        experiment_logger(rootdir='../tf_results', logdir_prefix=logdir_prefix, exp_config=exp_config, model=model,
-                          training_df=training_df,
-                          test_df=test_df)
+        logger.log_model_training(
+            exp_config=exp_config,
+            logdir_prefix=logdir_prefix,
+            model=model,
+            training_df=training_df,
+            test_df=test_df
+        )