KoRWKV-6B / tokenizer_config.json
beomi's picture
Add 6B model, trained 26B tokens
e453658
raw
history blame
236 Bytes
{
"bos_token": "<|sep|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<|endoftext|>",
"tokenizer_class": "PreTrainedTokenizerFast"
}