{
  "tokenizers": [
    {
      "name": "sp_bpe_8192",
      "dataset_suffix": "sp8192",
      "vocab_size": 8192,
      "tokenizer_train_docs": 50000
    }
  ]
}