| { | |
| "tokenizer_class": "Telechat2Tokenizer", | |
| "auto_map": { | |
| "AutoTokenizer": [ | |
| "tokenization_telechat2.Telechat2Tokenizer", | |
| null | |
| ] | |
| }, | |
| "added_tokens_decoder": { | |
| "1": { | |
| "content": "<_start>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "2": { | |
| "content": "<_end>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "3": { | |
| "content": "<_pad>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "4": { | |
| "content": "<_user>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "5": { | |
| "content": "<_bot>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| } | |
| }, | |
| "additional_special_tokens": [ | |
| "<_start>", | |
| "<_end>", | |
| "<_pad>", | |
| "<_user>", | |
| "<_bot>" | |
| ], | |
| "add_bos_token": false, | |
| "add_eos_token": false, | |
| "use_fast": false, | |
| "clean_up_tokenization_spaces": false, | |
| "split_special_tokens": false, | |
| "model_max_length": 100000000, | |
| "sp_model_kwargs": {}, | |
| "bos_token": "<_start>", | |
| "eos_token": "<_end>", | |
| "pad_token": "<_pad>", | |
| "chat_template": "<_start>{%- for message in messages %}{% if message.role == 'user' %}{{'<_user>'+message.content}}{% endif %}{% if message.role == 'bot' %}{{'<_bot>'+message.content+'<_end>'}}{% endif %}{% endfor %}{% if add_generation_prompt %}<_bot>{% endif %}" | |
| } | |