tokenizer_config.json

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