Unverified Commit 967b7ddc authored by Phil Wang's avatar Phil Wang Committed by GitHub
Browse files

Merge pull request #19 from eonglints/audio_selection_padding 

Random audio selection and padding as required
parents dfd4b803 836fa427

audiolm_pytorch/data.py

+9 −0
Original line number Diff line number Diff line
@@ -51,6 +51,15 @@ class SoundDataset(Dataset): 
        file = self.files[idx]



        data, sample_hz = torchaudio.load(file)

        

        if data.size(1) >= self.max_length:

            max_start = data.size(1) - self.max_length

            start = torch.randint(0, max_start, (1, ))

            data = data[:, start:start + self.max_length]



        else:

            data = torch.nn.functional.pad(data, (self.max_length - data.size(1), 0), 'constant')

        

        data = rearrange(data, '1 ... -> ...')



        num_outputs = len(self.target_sample_hz)