refactoring

nnet/xvector.py

@@ -268,31 +268,24 @@ def new_test_metrics(model,
     Returns:
         [type]: [description]
     """
-    # TODO modifier les parametres pour utiliser le dataset_description a la place de :
-
-    #idmap_test_filename,
-    #ndx_test_filename,
-    #key_test_filename,
-    #data_root_name,
-
     transform_pipeline = dict()
 
-    xv_stat = extract_embeddings(idmap_name=data_opts["idmap_test_filename"],
+    xv_stat = extract_embeddings(idmap_name=data_opts["test"]["idmap"],
                                  model_filename=model,
-                                 data_root_name=data_opts["data_root_name"],
+                                 data_root_name=data_opts["test"]["data_path"],
                                  device=device,
                                  loss=model.loss,
                                  transform_pipeline=transform_pipeline,
-                                 num_thread=train_opts["num_thread"],
+                                 num_thread=train_opts["num_cpu"],
                                  mixed_precision=train_opts["mixed_precision"])
 
     tar, non = cosine_scoring(xv_stat,
                               xv_stat,
-                              Ndx(data_opts["ndx_test_filename"]),
+                              Ndx(data_opts["test"]["ndx"]),
                               wccn=None,
                               check_missing=True,
                               device=device
-                              ).get_tar_non(Key(data_opts["key_test_filename"]))
+                              ).get_tar_non(Key(data_opts["test"]["key"]))
 
     #test_eer = eer(numpy.array(non).astype(numpy.double), numpy.array(tar).astype(numpy.double))
     pmiss, pfa = rocch(tar, non)
@@ -314,9 +307,10 @@ def save_checkpoint(state, is_best, filename='checkpoint.pth.tar', best_filename
         shutil.copyfile(filename, best_filename)
 
 
-
 class TrainingMonitor():
+    """
 
+    """
     def __init__(self,
                  output_file,
                  log_interval=10,
@@ -326,7 +320,7 @@ class TrainingMonitor():
                  best_eer=100,
                  compute_test_eer=False
                  ):
-        # Stocker plutot des listes pour conserver l'historique complet
+
         self.current_epoch = 0
         self.log_interval = log_interval
         self.init_patience = patience
@@ -347,13 +341,9 @@ class TrainingMonitor():
         self.is_best = True
 
         # Initialize the logger
-        logging.basicConfig(level=logging.DEBUG,
-                            format='%(asctime)s %(name)-12s %(levelname)-8s %(message)s',
-                            datefmt='%m-%d %H:%M',
-                            filename=output_file,
-                            filemode='w')
-
-        self.logger = logging.getLogger('monitoring')
+        logging_format = '%(asctime)-15s-8s %(message)s'
+        logging.basicConfig(level=logging.DEBUG, format=logging_format, datefmt='%m-%d %H:%M')
+        self.logger = logging.getLogger('Monitoring')
         self.logger.setLevel(logging.INFO)
         # create file handler which logs even debug messages
         fh = logging.FileHandler(output_file)
@@ -1002,9 +992,9 @@ def update_training_dictionary(dataset_description,
 
 def get_network(model_opts):
     """
-    
-    :param model_opts: 
-    :return: 
+
+    :param model_opts:
+    :return:
     """
 
     if model_opts["model_type"] in ["xvector", "rawnet2", "resnet34", "fastresnet34"]:
@@ -2008,8 +1998,6 @@ def extract_embeddings(idmap_name,
     else:
         min_duration = (model.module.context_size() - 1) * frame_shift + frame_duration
         model_cs = model.module.context_size()
-            
-
 
     # Create dataset to load the data
     dataset = IdMapSet(idmap_name=idmap_name,
@@ -2042,7 +2030,7 @@ def extract_embeddings(idmap_name,
             if name != 'bias':
                 name = name + '.weight'
             emb_size = model.module.before_speaker_embedding.state_dict()[name].shape[0]
-   
+
 
         # Create the StatServer
         embeddings = StatServer()