tokenizer_config.json

{"target_lang": "isl", "source_lang": "ita"}