tokenizer_config.json

{"tokenizer_class": "PreTrainedTokenizer"}