fingerprinted now should work correctly

audio_toolset/audio_io.py

@@ -125,4 +125,5 @@ class MelToImage(object):
         img = scale_minmax(mels, 0, 255).astype(np.uint8)
         img = np.flip(img, axis=0)  # put low frequencies at the bottom in image
         img = 255 - img  # invert. make black==more energy
+        img = img.astype(np.float32)
         return img

audio_toolset/mel_augmentation.py

@@ -59,9 +59,9 @@ class ShiftTime(object):
             # Set to silence for heading/ tailing
             shift = int(shift)
             if shift > 0:
-                augmented_data[:shift] = 0
+                augmented_data[:, :shift] = 0
             else:
-                augmented_data[shift:] = 0
+                augmented_data[:, shift:] = 0
             return augmented_data
         else:
             return x

utils/transforms.py

@@ -4,5 +4,6 @@ from torchvision.transforms import ToTensor as TorchVisionToTensor
 class ToTensor(TorchVisionToTensor):
 
     def __call__(self, pic):
+        # Make it float .float() == 32bit
         tensor = super(ToTensor, self).__call__(pic).float()
         return tensor