tokenizer_config.json

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