spectra-tokenizer-prec-training-r1 / tokenizer_config.json
imgremlin's picture
Upload tokenizer
35137d9
raw
history blame
207 Bytes
{
"bos_token": "<bos>",
"clean_up_tokenization_spaces": true,
"eos_token": "<eos>",
"model_max_length": 512,
"pad_token": "<pad>",
"tokenizer_class": "SpectraTokenizer",
"unk_token": "<unk>"
}