tokenizer_config.json

{"target_lang": "mfs", "source_lang": "es"}