tokenizer_config.json

{"target_lang": "mkd", "source_lang": "spa"}