simcse-ja-bert-base-clcmlp / tokenizer_config.json
akiFQC's picture
update tokenizer
63782c6
raw
history blame
456 Bytes
{
"do_lower_case": false,
"do_subword_tokenize": true,
"do_word_tokenize": true,
"mask_token": "[MASK]",
"mecab_kwargs": {
"mecab_dic": "unidic_lite"
},
"model_max_length": 512,
"pad_token": "[PAD]",
"cls_token": "[CLS]",
"sep_token": "[SEP]",
"special_tokens_map_file": null,
"name_or_path": "cl-tohoku/bert-base-japanese-v2",
"subword_tokenizer_type": "wordpiece",
"unk_token": "[UNK]",
"word_tokenizer_type": "mecab"
}