tokenizer_config.json

{"target_lang": "zls", "source_lang": "eng"}