diff --git a/codes/data/audio/unsupervised_audio_dataset.py b/codes/data/audio/unsupervised_audio_dataset.py index d764a83d..6ab7cf9c 100644 --- a/codes/data/audio/unsupervised_audio_dataset.py +++ b/codes/data/audio/unsupervised_audio_dataset.py @@ -136,12 +136,12 @@ class UnsupervisedAudioDataset(torch.utils.data.Dataset): if __name__ == '__main__': params = { 'mode': 'unsupervised_audio', - 'path': ['Z:\\split\\cleaned\\books0'], + 'path': ['Z:\\split\\cleaned\\books0', 'Z:\\split\\cleaned\\books2'], 'cache_path': 'E:\\audio\\remote-cache.pth', 'sampling_rate': 22050, 'pad_to_seconds': 5, 'phase': 'train', - 'n_workers': 0, + 'n_workers': 4, 'batch_size': 16, 'extra_samples': 4, } diff --git a/codes/scripts/audio/preparation/split_on_silence.py b/codes/scripts/audio/preparation/split_on_silence.py index 38b7859a..7f057a24 100644 --- a/codes/scripts/audio/preparation/split_on_silence.py +++ b/codes/scripts/audio/preparation/split_on_silence.py @@ -19,7 +19,7 @@ def main(): maximum_duration = 20 files = find_audio_files(args.path, include_nonwav=True) for e, wav_file in enumerate(tqdm(files)): - #if e < 1326: + #if e < 1459: # continue print(f"Processing {wav_file}..") outdir = os.path.join(args.out, f'{e}_{os.path.basename(wav_file[:-4])}').replace('.', '').strip()