Transformers
FlashFormerv1 / tokenizer_config.json
LastTransformer's picture
FIX: Complete compilation of 32k vocabulary into tokenizer.json
f483284 verified
Raw
History Blame Contribute Delete
412 Bytes
{
"add_prefix_space": null,
"backend": "tokenizers",
"bos_token": "<s>",
"clean_up_tokenization_spaces": false,
"eos_token": "</s>",
"extra_special_tokens": [
"<s>",
"</s>",
"<pad>"
],
"is_local": true,
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"tokenizer_class": "LlamaTokenizer",
"unk_token": "<unk>",
"use_default_system_prompt": false
}