Final PyTorch 1.8 release - freeze requirements - remove class_dict.k…

…eys() from kws20 (#324)

* Final PyTorch 1.8 release - freeze requirements, update README
* Remove .keys() from kws20.py
* Remove noise_type argument from kws20 get_datasets()

README.md

@@ -1,8 +1,8 @@
 # ADI MAX78000/MAX78002 Model Training and Synthesis
 
-June 25, 2024
+July 2, 2024
 
-**Note: This branch is compatible with PyTorch 1.8. Please go to the “pytorch-2” branch for PyTorch 2.3 compatibility.**
+**Note: This archived branch is compatible with PyTorch 1.8. Please go to the “develop” branch for PyTorch 2 compatibility.**
 
 ADI’s MAX78000/MAX78002 project is comprised of five repositories:
 

datasets/kws20.py

@@ -487,15 +487,13 @@ def __filter_dtype(self):
     def __filter_classes(self):
         initial_new_class_label = len(self.class_dict)
         new_class_label = initial_new_class_label
-        self.new_class_dict = {}
         for c in self.classes:
             if c not in self.class_dict:
                 if c == '_unknown_':
                     continue
                 raise ValueError(f'Class {c} not found in data')
             num_elems = (self.targets == self.class_dict[c]).cpu().sum()
             print(f'Class {c} (# {self.class_dict[c]}): {num_elems} elements')
-            self.new_class_dict[c] = new_class_label
             self.targets[(self.targets == self.class_dict[c])] = new_class_label
             new_class_label += 1
 
@@ -504,10 +502,6 @@ def __filter_classes(self):
         self.targets[(self.targets < initial_new_class_label)] = new_class_label
         self.targets -= initial_new_class_label
 
-        self.new_class_dict = {c: self.new_class_dict[c] - initial_new_class_label
-                               for c in self.new_class_dict.keys()}
-        self.new_class_dict['_unknown_'] = len(self.new_class_dict)
-
     def __filter_librispeech(self):
 
         print('Filtering out librispeech elements...')
@@ -1002,7 +996,6 @@ def KWS_35_get_unquantized_datasets(data, load_train=True, load_test=True):
 
 def KWS_20_msnoise_mixed_get_datasets(data, load_train=True, load_test=True,
                                       apply_prob=0.8, snr_range=(-5, 10),
-                                      noise_type=MSnoise.class_dict.keys(),
                                       desired_probs=None):
     """
     Returns the KWS dataset mixed with MSnoise dataset. Only training set will be mixed
@@ -1016,6 +1009,8 @@ def KWS_20_msnoise_mixed_get_datasets(data, load_train=True, load_test=True,
     noise_type --> All noise types in the noise dataset.
     """
 
+    noise_type = MSnoise.class_dict
+
     if len(snr_range) > 1:
         snr_range = range(snr_range[0], snr_range[1])
     else:
@@ -1058,14 +1053,15 @@ def KWS_12_benchmark_get_datasets(data, load_train=True, load_test=True):
 
 def MixedKWS_20_get_datasets_10dB(data, load_train=True, load_test=True,
                                   apply_prob=1, snr_range=tuple([10]),
-                                  noise_type=MSnoise.class_dict.keys(),
                                   desired_probs=None):
     """
     Returns the mixed KWS dataset with MSnoise dataset under 10 dB SNR using signalmixer
     data loader. All of the training and test data will be augmented with
     additional noise.
     """
 
+    noise_type = MSnoise.class_dict
+
     if len(snr_range) > 1:
         snr_range = range(snr_range[0], snr_range[1])
     else:

requirements.txt

@@ -1,26 +1,26 @@
 torch==1.8.1
 torchaudio==0.8.1
 torchvision==0.9.1
-GitPython>=3.1.18
-Pillow>=7
-PyYAML>=5.1.1
-albumentations>=1.3.0
+GitPython==3.1.43
+Pillow==10.4.0
+PyYAML==6.0.1
+albumentations==1.3.1
 faiss-cpu==1.7.4
-batch-face>=1.4.0
-h5py>=3.7.0
+batch-face==1.4.0
+h5py==3.11.0
 kornia==0.6.8
-librosa>=0.7.2
-numba<0.50.0
-numpy>=1.22,<1.23
-opencv-python>=4.4.0
-protobuf>=3.20.1,<4.0
+librosa==0.9.2
+numba==0.49.1
+numpy==1.22.4
+opencv-python==4.10.0.84
+protobuf==3.20.3
 pycocotools==2.0.7
 pyffmpeg==2.0
 pytorch-metric-learning==2.3.0
-pytube>=12.1.3
-scipy>=1.3.0
-shap>=0.34.0
-tensorboard>=2.9.0,<2.10.0
-tk>=0.1.0
+pytube==15.0.0
+scipy==1.10.1
+shap==0.44.1
+tensorboard==2.9.0
+tk==0.1.0
 torchmetrics==0.6.0
 -e distiller