Files
llama3-8b-full-pretrain-was…/tokenizer_config.json

20 lines
459 B
JSON
Raw Permalink Normal View History

{
"backend": "tokenizers",
"bos_token": "<|begin_of_text|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|eot_id|>",
"extra_special_tokens": [
"<|eom_id|>"
],
"is_local": true,
"model_input_names": [
"input_ids",
"attention_mask"
],
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<|eot_id|>",
"padding_side": "right",
"split_special_tokens": false,
"tokenizer_class": "TokenizersBackend"
}