Merge branch 'master' of gitlab.lrz.de:mobile-ifi/inter_challenge_2020

_paramters.py

@@ -28,6 +28,7 @@ main_arg_parser.add_argument("--data_sr", type=int, default=16000, help="")
 main_arg_parser.add_argument("--data_hop_length", type=int, default=256, help="")
 main_arg_parser.add_argument("--data_n_fft", type=int, default=512, help="")
 main_arg_parser.add_argument("--data_mixup", type=strtobool, default=False, help="")
+main_arg_parser.add_argument("--data_stretch", type=strtobool, default=False, help="")
 
 # Transformation Parameters
 main_arg_parser.add_argument("--data_loudness_ratio", type=float, default=0, help="")  # 0.4

datasets/binar_masks.py

@@ -19,7 +19,7 @@ class BinaryMasksDataset(Dataset):
     def sample_shape(self):
         return self[0][0].shape
 
-    def __init__(self, data_root, setting, mel_transforms, transforms=None, mixup=False, stretch_dataset=True):
+    def __init__(self, data_root, setting, mel_transforms, transforms=None, mixup=False, stretch_dataset=False):
         self.stretch = stretch_dataset
         assert isinstance(setting, str), f'Setting has to be a string, but was: {type(setting)}.'
         assert setting in V.DATA_OPTIONS, f'Setting must match one of: {V.DATA_OPTIONS}.'

models/residual_conv_classifier.py

@@ -39,15 +39,18 @@ class ResidualConvClassifier(BinaryMaskDatasetMixin,
         self.conv_list.append(ConvModule(last_shape, self.conv_filters[0], (k, k), conv_stride=(2, 2), conv_padding=1,
                                          **self.params.module_kwargs))
         last_shape = self.conv_list[-1].shape
-        for filters in self.conv_filters:
+        for idx in range(len(self.conv_filters)):
-            conv_module_params.update(conv_filters=filters)
+            conv_module_params.update(conv_filters=self.conv_filters[idx])
             self.conv_list.append(ResidualModule(last_shape, ConvModule, 3, **conv_module_params))
             last_shape = self.conv_list[-1].shape
-            self.conv_list.append(ConvModule(last_shape, filters, (k, k), conv_stride=(2, 2), conv_padding=2,
+            try:
-                                             **self.params.module_kwargs))
+                self.conv_list.append(ConvModule(last_shape, self.conv_filters[idx+1], (k, k), conv_stride=(2, 2), conv_padding=2,
-            for param in self.conv_list[-1].parameters():
+                                                 **self.params.module_kwargs))
-                param.requires_grad = False
+                for param in self.conv_list[-1].parameters():
-            last_shape = self.conv_list[-1].shape
+                    param.requires_grad = False
+                last_shape = self.conv_list[-1].shape
+            except IndexError:
+                pass
 
         self.full_1 = LinearModule(self.conv_list[-1].shape, self.params.lat_dim, **self.params.module_kwargs)
         self.full_2 = LinearModule(self.full_1.shape, self.full_1.shape * 2, **self.params.module_kwargs)

util/module_mixins.py

@@ -142,7 +142,7 @@ class BinaryMaskDatasetMixin:
             **dict(
                 # TRAIN DATASET
                 train_dataset=BinaryMasksDataset(self.params.root, setting=V.DATA_OPTIONS.train,
-                                                 mixup=self.params.mixup,
+                                                 mixup=self.params.mixup, stretch_dataset=self.params.stretch,
                                                  mel_transforms=mel_transforms_train, transforms=aug_transforms),
                 # VALIDATION DATASET
                 val_train_dataset=BinaryMasksDataset(self.params.root, setting=V.DATA_OPTIONS.train,