funasr/utils/load_utils.py
@@ -43,7 +43,7 @@ elif isinstance(data_or_path_or_list, str) and data_type == "text" and tokenizer is not None: data_or_path_or_list = tokenizer.encode(data_or_path_or_list) elif isinstance(data_or_path_or_list, np.ndarray): # audio sample point data_or_path_or_list = np.squeeze(data_or_path_or_list) # [n_samples,] data_or_path_or_list = torch.from_numpy(data_or_path_or_list).squeeze() # [n_samples,] else: pass # print(f"unsupport data type: {data_or_path_or_list}, return raw data")
