{"image_url":"https://cdn-ak.f.st-hatena.com/images/fotolife/e/e4exp/20211224/20211224151250.png","version":"1.0","published":"2021-12-24 15:15:46","height":"190","url":"https://e4exp.hatenablog.com/entry/2021/12/24/151546","categories":["DeepLearning","Pre-Training","Vision-Language","Transformer"],"provider_url":"https://hatena.blog","description":"Florence: A New Foundation Model for Computer Vision 2021/11/22 https://arxiv.org/abs/2111.11432 Fig.2 Overview of building Florence \u753b\u50cf\u30c9\u30e1\u30a4\u30f3\u3067\u591a\u69d8\u306a\u4e0b\u6d41\u30bf\u30b9\u30af(\u5206\u985e\u3001\u691c\u7d22\u3001\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u691c\u51fa\u3001VQA\u3001\u753b\u50cf\u30ad\u30e3\u30d7\u30b7\u30e7\u30f3\u3001\u30d3\u30c7\u30aa\u691c\u7d22\u3001\u30a2\u30af\u30b7\u30e7\u30f3\u8a8d\u8b58\u306a\u3069)\u306b\u9069\u7528\u3067\u304d\u308b\u6c4e\u7528\u7684\u306a\u4e8b\u524d\u5b66\u7fd2\u30e2\u30c7\u30ebFlorence\u3092\u63d0\u6848 \u30e2\u30c7\u30eb\u306f\u968e\u5c64\u578bvision transformer(CoSwin)\u306e\u753b\u50cfencoder\u3068Transformer \u8a00\u8a9eencoder\u3067image-text\u9593\u306e\u2026","type":"rich","width":"100%","author_name":"e4exp","blog_title":"\u5b66\u3093\u3060\u3053\u3068\u30e1\u30e2","blog_url":"https://e4exp.hatenablog.com/","html":"<iframe src=\"https://hatenablog-parts.com/embed?url=https%3A%2F%2Fe4exp.hatenablog.com%2Fentry%2F2021%2F12%2F24%2F151546\" title=\"vision\u5206\u91ce\u3067\u591a\u69d8\u306a\u4e0b\u6d41\u30bf\u30b9\u30af\u306b\u9069\u7528\u3067\u304d\u308b\u57fa\u790e\u30e2\u30c7\u30ebFlorence - \u5b66\u3093\u3060\u3053\u3068\u30e1\u30e2\" class=\"embed-card embed-blogcard\" scrolling=\"no\" frameborder=\"0\" style=\"display: block; width: 100%; height: 190px; max-width: 500px; margin: 10px 0px;\"></iframe>","provider_name":"Hatena Blog","title":"vision\u5206\u91ce\u3067\u591a\u69d8\u306a\u4e0b\u6d41\u30bf\u30b9\u30af\u306b\u9069\u7528\u3067\u304d\u308b\u57fa\u790e\u30e2\u30c7\u30ebFlorence","author_url":"https://blog.hatena.ne.jp/e4exp/"}