tokenizer_config.json

{"target_lang": "swe", "source_lang": "zho"}