tokenizer_config.json

{"target_lang": "tw", "source_lang": "sv"}