ngme-babylm-100M / tokenizer_config.json
PatrickHaller's picture
Upload tokenizer
d8d7eca
raw
history blame contribute delete
321 Bytes
{
"auto_map": {
"AutoTokenizer": [
"tokenization_ngme.NGMETokenizer",
null
]
},
"clean_up_tokenization_spaces": true,
"eos_token": "\n",
"model_max_length": 1000000000000000019884624838656,
"ngram": null,
"pad_token": "\n",
"tokenizer_class": "NGMETokenizer",
"unk_token": "<unk>"
}