tokenizer_config.json

{"target_lang": "transformers", "source_lang": "", "separate_vocabs": false}