tokenizer_config.json

{
    "do_lower_case": false,
    "bos_token": "<s>",
    "eos_token": "</s>",
    "unk_token": "<unk>",
    "sep_token": "</s>",
    "pad_token": "<pad>",
    "cls_token": "<s>",
    "mask_token": "<mask>",
    "split_by_punct": false,
    "sp_model_kwargs": {},
    "special_tokens_map_file": null,
    "name_or_path": "/cognitive_comp/gaoxinyu/pretrained_model/deberta-base-sp",
    "tokenizer_class": "DebertaV2Tokenizer"
}