SpliNet / tokenizer_config.json
Angshul's picture
Upload repaired SpliNet 2B-token pretrained model
4ec5e47 verified
Raw History Blame Contribute Delete
370 Bytes
{
"tokenizer_class": "SpliNetTokenizer",
"auto_map": {
"AutoTokenizer": [
"tokenization_splinet.SpliNetTokenizer",
null
]
},
"model_max_length": 512,
"do_lower_case": true,
"unk_token": "<unk>",
"bos_token": "<s>",
"eos_token": "</s>",
"pad_token": "<pad>",
"cls_token": "<cls>",
"sep_token": "<sep>",
"mask_token": "<mask>"
}