DeepSeekMath-Base-SFT-Step-DPO / tokenizer_config.json
xinlai's picture
upload model
07ddda3
{
"add_bos_token": true,
"add_eos_token": false,
"added_tokens_decoder": {
"100000": {
"content": "<|begin▁of▁sentence|>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
},
"100001": {
"content": "<|end▁of▁sentence|>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
},
"100002": {
"content": "[PAD]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"bos_token": "<|begin▁of▁sentence|>",
"clean_up_tokenization_spaces": false,
"eos_token": "<|end▁of▁sentence|>",
"legacy": true,
"max_length": 512,
"model_max_length": 512,
"pad_to_multiple_of": null,
"pad_token": "[PAD]",
"pad_token_type_id": 0,
"padding_side": "right",
"sp_model_kwargs": {},
"stride": 0,
"tokenizer_class": "LlamaTokenizer",
"truncation_side": "left",
"truncation_strategy": "longest_first",
"unk_token": null,
"use_default_system_prompt": false
}