tokenizer_config.json

{"target_lang": "ja", "source_lang": "en-de-fr-it-es-ru-uk"}