tokenizer_config.json

{
  "bos_token": "<s>",
  "clean_up_tokenization_spaces": true,
  "do_lower_case": false,
  "langs": [
    "en",
    "th"
  ],
  "model_max_length": 1024,
  "pad_token": "<pad>",
  "sep_token": "</s>",
  "tokenizer_class": "VISTECFSMTTokenizer",
  "unk_token": "<unk>"
}