roberta-base-latin-ud-goeswith / tokenizer_config.json
KoichiYasuoka's picture
initial release
d8ebfde
{
"add_prefix_space": true,
"bos_token": "<s>",
"cls_token": "<s>",
"eos_token": "</s>",
"errors": "replace",
"keep_accents": true,
"mask_token": "<mask>",
"model_max_length": 514,
"pad_token": "<pad>",
"sep_token": "</s>",
"tokenizer_class": "RobertaTokenizerFast",
"trim_offsets": true,
"unk_token": "<unk>"
}