tokenizer_config.json

{"target_lang": "fin", "source_lang": "kor"}