mispeech/MECAT-QA
收藏官方服务：
资源简介：
MECAT（多专家链音频任务）是一个建立在大量数据之上的全面基准测试，旨在通过两个核心任务——音频标注和音频问答，评估机器对音频内容的理解。
MECAT (Multi-Expert Chain for Audio Tasks) is a comprehensive benchmark constructed on large-scale data to evaluate machine understanding of audio content through two core tasks: Audio Captioning and Audio Question Answering.
提供机构：
mispeech
