update

shixian.shi

2023-06-27 25de54910e7b48552bdba2dd655fbcd64a07668e

 funasr/datasets/large_datasets/utils/tokenize.py

@@ -63,7 +63,6 @@
                if _find != -1:
                    # _find = text[:_find].count(" ")  # bpe sometimes
                    pre_index = [_find, _find + max(hw.count(" "), 1)]
                    # import pdb; pdb.set_trace()
                    break
        hotword_indxs = sample_hotword(length, **hw_config, pre_index=pre_index)
        data['hotword_indxs'] = hotword_indxs