debug

nnet/xsets.py

@@ -293,8 +293,8 @@ class IdMapSet(Dataset):
                  idmap_name,
                  data_path,
                  file_extension,
-                 transform_pipeline={},
-                 sliding_window=True,
+                 transform_pipeline="",
+                 sliding_window=False,
                  window_len=24000,
                  window_shift=8000,
                  sample_rate=16000,
@@ -358,9 +358,9 @@ class IdMapSet(Dataset):
                 start = max(0, int(middle - (self.min_sample_nb / 2)))
                 duration = self.min_sample_nb
 
-        speech, speech_fs = torchaudio.load(f"{self.data_path}/{self.idmap.rightids[index]}.{self.file_extension}",
-                                            frame_offset=start,
-                                            num_frames=duration)
+            speech, speech_fs = torchaudio.load(f"{self.data_path}/{self.idmap.rightids[index]}.{self.file_extension}",
+                                                frame_offset=start,
+                                                num_frames=duration)
 
         speech += 10e-6 * torch.randn(speech.shape)
 

nnet/xvector.py

@@ -226,7 +226,7 @@ def test_metrics(model,
     key_test_filename = 'h5f/key_test.h5'
     data_root_name='/lium/scratch/larcher/voxceleb1/test/wav'
 
-    transform_pipeline = dict()
+    transform_pipeline = ""
 
     xv_stat = extract_embeddings(idmap_name=idmap_test_filename,
                                  model_filename=model,
@@ -837,9 +837,6 @@ def xtrain(speaker_number,
             new_model_dict.update(pretrained_dict)
             model.load_state_dict(new_model_dict)
 
-            print("Modifiy margin: 0.4")
-            model.after_speaker_embedding.m = 0.4
-
 
         # Freeze required layers
         for name, param in model.named_parameters():
@@ -1201,7 +1198,6 @@ def cross_validation(model, validation_loader, device, validation_shape, mask, m
     :return:
     """
     model.eval()
-    print("In cross validation")
     if isinstance(model, Xtractor):
         loss_criteria = model.loss
     else:
@@ -1214,9 +1210,7 @@ def cross_validation(model, validation_loader, device, validation_shape, mask, m
     classes = torch.zeros([validation_shape[0]])
     with torch.no_grad():
         #for batch_idx, (data, target) in enumerate(tqdm.tqdm(validation_loader, desc='validation compute', mininterval=1)):
-        print("In cross validation 2")
         for batch_idx, (data, target) in enumerate(validation_loader):
-            print("In cross validation 3")
             batch_size = target.shape[0]
             target = target.squeeze().to(device)
             data = data.squeeze().to(device)
@@ -1254,7 +1248,7 @@ def extract_embeddings(idmap_name,
                        data_root_name,
                        device,
                        file_extension="wav",
-                       transform_pipeline={},
+                       transform_pipeline="",
                        frame_shift=0.01,
                        frame_duration=0.025,
                        extract_after_pooling=False,
@@ -1306,7 +1300,6 @@ def extract_embeddings(idmap_name,
                        data_path=data_root_name,
                        file_extension=file_extension,
                        transform_pipeline=transform_pipeline,
-                       frame_rate=int(1. / frame_shift),
                        min_duration=(model_cs + 2) * frame_shift * 2
                        )
 
@@ -1602,7 +1595,7 @@ def extract_sliding_embedding(idmap_name,
                     segset += [seg,] *  data.shape[0]
                     starts += [numpy.arange(start, start + embeddings.shape[0] * window_shift , window_shift),]
 
-REPRENDRE ICI
+        #REPRENDRE ICI
 
         # Create the StatServer
         embeddings = StatServer()