| | |
| | | preprocessor = LMPreprocessor( |
| | | train=False, |
| | | token_type=speechtext2timestamp.tp_train_args.token_type, |
| | | token_list=speechtext2timestamp.tp_train_args, |
| | | token_list=speechtext2timestamp.tp_train_args.token_list, |
| | | bpemodel=None, |
| | | text_cleaner=None, |
| | | g2p_type=None, |
| | |
| | | batch_size=batch_size, |
| | | key_file=key_file, |
| | | num_workers=num_workers, |
| | | preprocess_fn=LMPreprocessor, |
| | | preprocess_fn=preprocessor, |
| | | collate_fn=ASRTask.build_collate_fn(speechtext2timestamp.tp_train_args, False), |
| | | allow_variable_data_keys=allow_variable_data_keys, |
| | | inference=True, |