tokenizer_config.json

{"target_lang": "ukr", "source_lang": "hbs"}