saigaTABv3 / tokenizer_config.json
marcus2000's picture
Upload 2 files
0cbc665 verified
raw
history blame contribute delete
278 Bytes
{
"tokenizer_class": "LlamaTokenizer",
"model_max_length": 4096,
"padding_side": "left",
"bos_token": "<s>",
"eos_token": "</s>",
"unk_token": "<unk>",
"clean_up_tokenization_spaces": false,
"special_tokens_map_file": "special_tokens_map.json"
}