TeleChat-12B-int8 / tokenizer_config.json
liuxz0801's picture
Update tokenizer_config.json
cafedaa verified
{
"name_or_path": "ChinaTelecom/telechat-12b",
"tokenizer_class": "TelechatTokenizer",
"auto_map": {
"AutoTokenizer": [
"tokenization_telechat.TelechatTokenizer",
null
]
},
"add_bos_token": false,
"add_eos_token": false,
"use_fast": false,
"clean_up_tokenization_spaces": false,
"eos_token": {
"__type": "AddedToken",
"content": "<_end>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": true
},
"model_max_length": 100000000,
"sp_model_kwargs": {},
"pad_token": {
"__type": "AddedToken",
"content": "<_pad>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": true
},
"unk_token": {
"__type": "AddedToken",
"content": "<_end>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": true
}
}