Prompting Visual-Language Models for Efficient Video Understanding2022-11-040 PRE AI DOI:10.1007/978-3-031-19833-5_7原文链接分享收藏摘要 En