tokenizer_config.json

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