lmms-lab/MixEval-X-audio2text

该数据集包含多个特征，如id、问题类型、输入文件、查询、选项、参考答案、基准名称和音频。音频特征具有192000的采样率。数据集分为两个部分：free_form和free_form_hard，分别包含962和505个例子，以及相应的文件大小。此外，还提供了数据集的下载大小和总大小。

This dataset includes multiple features such as id, problem_type, input_file, query, options, reference_answer, benchmark_name, and audio. The audio feature has a sampling rate of 192000. The dataset is divided into two parts: free_form and free_form_hard, containing 962 and 505 examples respectively, along with their respective file sizes. Additionally, the download size and total size of the dataset are provided.