tokenizer_config.json

{"target_lang": "bul", "source_lang": "epo"}