tokenizer_config.json

{"target_lang": "rw", "source_lang": "fr"}