tokenizer_config.json

{"sep_token": "<sep>", "pad_token": "<pad>", "cls_token": "<cls>", "mask_token": "<mask>", "unk_token": "<unk>", "tokenizer_class": "PreTrainedTokenizerFast"}