已收录文章摘要，完整内容可前往原文阅读。
原标题：Who Needs Labels? Adapting Vision Foundation Models With the Metadata You Already Have下载：https:/