deberta-v2-base-japanese / tokenizer_config.json
nobu-g's picture
first commit
11debab
raw
history blame
400 Bytes
{
"bos_token": "[CLS]",
"cls_token": "[CLS]",
"do_lower_case": false,
"eos_token": "[SEP]",
"keep_accents": true,
"mask_token": "[MASK]",
"name_or_path": "data/deberta_v2_tokenizer",
"pad_token": "[PAD]",
"sep_token": "[SEP]",
"sp_model_kwargs": {},
"special_tokens_map_file": null,
"split_by_punct": false,
"tokenizer_class": "DebertaV2Tokenizer",
"unk_token": "[UNK]"
}