funasr/models/frontend/wav_frontend.py
@@ -145,9 +145,12 @@ feats_lens.append(feat_length) feats_lens = torch.as_tensor(feats_lens) feats_pad = pad_sequence(feats, batch_first=True, padding_value=0.0) if batch_size == 1: feats_pad = feats[0][None, :, :] else: feats_pad = pad_sequence(feats, batch_first=True, padding_value=0.0) return feats_pad, feats_lens def forward_fbank(
