liz-nojaloli-nxja-ja / tokenizer_config.json
ebisuke's picture
initial commit
ce42fdb
raw
history blame contribute delete
276 Bytes
{
"bos_token": "<|startoftext|>",
"clean_up_tokenization_spaces": true,
"do_clean_text": false,
"eos_token": "<|endoftext|>",
"model_max_length": 2048,
"pad_token": "<|endoftext|>",
"tokenizer_class": "GPTNeoXJapaneseTokenizer",
"unk_token": "<|endoftext|>"
}