tokenizer_config.json

{
    "tokenizer_type": "AATokenizer",
    "vocabulary_path": "data/vocabulary/aa.txt"
}