tokenizer_config.json

{"target_lang": "en+el+es+fi", "source_lang": "en+el+es+fi"}