tokenizer_config.json

{
  "bos_token": "<s>",
  "cls_token": "<s>",
  "eos_token": "</s>",
  "mask_token": "<mask>",
  "model_max_length": 512,
  "name_or_path": "data/plms/legal-latvian-roberta-base",
  "pad_token": "<pad>",
  "sep_token": "</s>",
  "special_tokens_map_file": "data/plms/legal-latvian-roberta-base_32k/special_tokens_map.json",
  "tokenizer_class": "PreTrainedTokenizerFast",
  "unk_token": "<unk>"
}