bert-base-japanese-char-v3 / tokenizer_config.json
cl-tohoku's picture
Add model files
ccddee5
raw
history blame contribute delete
251 Bytes
{
"tokenizer_class": "BertJapaneseTokenizer",
"model_max_length": 512,
"do_lower_case": false,
"word_tokenizer_type": "mecab",
"subword_tokenizer_type": "character",
"mecab_kwargs": {
"mecab_dic": "unidic_lite"
}
}