runtime/python/libtorch/funasr_torch/sensevoice_bin.py
@@ -111,10 +111,8 @@ mask = yseq != self.blank_id token_int = yseq[mask].tolist() if tokenizer is not None: asr_res.append(self.tokenizer.decode(token_int)) else: asr_res.append(token_int) asr_res.append(self.tokenizer.decode(token_int)) return asr_res def load_data(self, wav_content: Union[str, np.ndarray, List[str]], fs: int = None) -> List:
