tokenizer_config.json

{"target_lang": "gmw", "source_lang": "gmw"}