llama-600M-v4-isolate / tokenizer_config.json
deqing's picture
Sync main at tokens-6.4B tokens
d068ff9 verified
{
"backend": "tokenizers",
"bos_token": "<|begin_of_text|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|end_of_text|>",
"is_local": false,
"model_input_names": [
"input_ids",
"attention_mask"
],
"model_max_length": 131072,
"tokenizer_class": "TokenizersBackend"
}