diff --git a/egs/librispeech/ASR/local/.prepare_vox.py.swp b/egs/librispeech/ASR/local/.prepare_vox.py.swp index c7939fc0b..cce5c3800 100644 Binary files a/egs/librispeech/ASR/local/.prepare_vox.py.swp and b/egs/librispeech/ASR/local/.prepare_vox.py.swp differ diff --git a/egs/librispeech/ASR/local/prepare_vox.py b/egs/librispeech/ASR/local/prepare_vox.py index ff1375758..d956a61b5 100755 --- a/egs/librispeech/ASR/local/prepare_vox.py +++ b/egs/librispeech/ASR/local/prepare_vox.py @@ -59,9 +59,10 @@ def prepare_vox( # ) manifests = {} - + subset = spk_id.split('_')[0] + spk_idx = spk_id.split('_')[1] #dataset_parts = ["train", "dev", "test"] - dataset_parts = [spk_id] + dataset_parts = [spk_idx] if output_dir is not None: output_dir = Path(output_dir) output_dir.mkdir(parents=True, exist_ok=True) @@ -81,7 +82,8 @@ def prepare_vox( #part_file_names = list(map(lambda x: x.strip('.wav'), os.listdir(part_path))) part_file_names = sorted(glob.glob(str(part_path)+'/*.wav')) part_file_names = [name.split('/')[-1].replace('.wav', '') for name in part_file_names] - txt_path = os.path.join(corpus_dir, f"../{spk_id}/texts") + txt_path = os.path.join(corpus_dir, f"../{spk_idx}_texts") + print(txt_path) futures = [] for trans_path in tqdm(