{"id":8336,"date":"2026-05-19T17:06:37","date_gmt":"2026-05-19T09:06:37","guid":{"rendered":"https:\/\/infernews.com\/blog\/towards-multimodal-skills-for-general-visual-agents\/"},"modified":"2026-05-19T17:06:37","modified_gmt":"2026-05-19T09:06:37","slug":"towards-multimodal-skills-for-general-visual-agents","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/towards-multimodal-skills-for-general-visual-agents\/","title":{"rendered":"MMSkills\uff1a\u5e6b\u8996\u89ba\u4ee3\u7406\u5b78\u8b58\u7747\u756b\u9762\u505a\u4e8b"},"content":{"rendered":"<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/05\/full_figure-f7cbee441fb4.jpg\" alt=\"MMSkills\"><\/figure>\n<p>MMSkills \u662f\u4e00\u500b\u70ba\u8996\u89ba\u4ee3\u7406\u800c\u8a2d\u7684\u6280\u80fd\u6846\u67b6\uff0c\u91cd\u9ede\u4e0d\u662f\u55ae\u9760\u6587\u5b57\u63d0\u793a\uff0c\u800c\u662f\u628a\u300c\u9ede\u505a\u4e00\u4ef6\u4e8b\u300d\u9023\u540c\u756b\u9762\u7dda\u7d22\u3001\u72c0\u614b\u63d0\u793a\u548c\u53ef\u53c3\u8003\u5f71\u50cf\u4e00\u4f75\u4fdd\u5b58\u3002\u5c0d\u4e00\u822c\u8b80\u8005\u4f86\u8aaa\uff0c\u53ef\u4ee5\u7406\u89e3\u6210\u66ff AI \u6e96\u5099\u4e00\u672c\u6703\u6309\u60c5\u5883\u7ffb\u95b1\u7684\u64cd\u4f5c\u624b\u518a\uff0c\u800c\u4e0d\u662f\u6bcf\u6b21\u7531\u96f6\u958b\u59cb\u4f30\u3002<\/p>\n<p>MMSkills \u5617\u8a66\u89e3\u6c7a\u8996\u89ba\u4ee3\u7406\u5728\u684c\u9762\u64cd\u4f5c\u6216\u904a\u6232\u5834\u666f\u4e2d\uff0c\u7d93\u5e38\u77e5\u9053\u76ee\u6a19\u537b\u672a\u5fc5\u8b58\u5f97\u5224\u65b7\u756b\u9762\u9032\u5ea6\u3001\u5931\u6557\u8de1\u8c61\uff0c\u6216\u8005\u4e0b\u4e00\u6b65\u61c9\u5426\u8f49\u62db\u3002MMSkills \u7684\u505a\u6cd5\u662f\u53ea\u4fdd\u7559\u8f15\u91cf\u63d0\u793a\uff0c\u7576\u5075\u6e2c\u5230\u7576\u524d\u72c0\u614b\u9069\u5408\u67d0\u500b\u6280\u80fd\u6642\uff0c\u624d\u81e8\u6642\u958b\u4e00\u500b\u5206\u652f\u53bb\u67e5\u770b\u8f03\u5b8c\u6574\u7684\u6b65\u9a5f\u8207\u8996\u89ba\u53c3\u8003\uff0c\u6e1b\u5c11\u4e3b\u4e0a\u4e0b\u6587\u8ca0\u64d4\u3002<\/p>\n<p>\u7559\u610f agent adapter \u5982\u4f55\u63a5\u5165\u73fe\u6709\u4ee3\u7406\u3002\u516c\u958b\u8cc7\u6599\u986f\u793a\uff0c\u5b83\u5df2\u63d0\u4f9b\u53ef\u641c\u5c0b\u7684\u591a\u9818\u57df\u6280\u80fd\u5eab\uff0c\u4e26\u652f\u63f4\u628a\u6280\u80fd\u5305\u63a5\u5230\u4e0d\u540c\u4ee3\u7406\u6d41\u7a0b\uff0c\u8f03\u9069\u5408\u5df2\u6709\u4ee3\u7406\u7cfb\u7d71\u3001\u60f3\u63d0\u5347\u7a69\u5b9a\u6027\u7684\u4eba\u4f7f\u7528\u3002<\/p>\n<ul>\n<li>\u628a\u6587\u5b57\u6b65\u9a5f\u3001\u72c0\u614b\u5361\u8207\u95dc\u9375\u756b\u9762\u7d50\u5408\uff0c\u4e0d\u53ea\u9760 prompt<\/li>\n<li>\u6280\u80fd\u53ef\u641c\u5c0b\u53ca\u91cd\u7528\uff0c\u73fe\u6709\u7db2\u7ad9\u986f\u793a\u6280\u80fd\u5eab\u898f\u6a21\u9054 515 \u500b<\/li>\n<li>\u63a1\u7528\u81e8\u6642\u5206\u652f\u8b80\u53d6\u6280\u80fd\uff0c\u8f03\u50cf\u6309\u9700\u8981\u67e5\u624b\u518a<\/li>\n<li>\u9069\u7528\u65bc GUI \u8207\u904a\u6232\u578b\u8996\u89ba\u4efb\u52d9\uff0c\u91cd\u9ede\u5728\u904b\u884c\u6642\u6c7a\u7b56<\/li>\n<li>\u53ef\u5c0d\u63a5 Codex\u3001OpenClaw\u3001Claude Code \u7b49\u4ee3\u7406<\/li>\n<\/ul>\n<p>\u9019\u500b\u9805\u76ee\u7684\u65b0\u610f\uff0c\u5728\u65bc\u5b83\u4e0d\u53ea\u7814\u7a76\u6a21\u578b\u672c\u8eab\uff0c\u800c\u662f\u628a\u5916\u90e8\u7d93\u9a57\u6574\u7406\u6210\u53ef\u91cd\u7528\u77e5\u8b58\u5305\uff0c\u518d\u5728\u63a8\u7406\u6642\u6309\u72c0\u614b\u8abf\u7528\u3002\u8ad6\u6587\u6458\u8981\u4ea6\u63d0\u5230\uff0c\u9019\u5957\u65b9\u6cd5\u5c0d\u524d\u6cbf\u6a21\u578b\u548c\u8f03\u5c0f\u578b\u591a\u6a21\u614b\u6a21\u578b\u90fd\u6709\u5e6b\u52a9\uff1b\u5982\u679c\u4f60\u6b63\u5728\u95dc\u6ce8\u96fb\u8166\u4ee3\u64cd\u4f5c\u3001\u81ea\u52d5\u5316\u6e2c\u8a66\u3001\u904a\u6232\u4ee3\u7406\uff0c\u6216\u8005\u60f3\u7814\u7a76 AI \u5982\u4f55\u300c\u7747\u756b\u9762\u518d\u884c\u52d5\u300d\uff0cMMSkills \u503c\u5f97\u7559\u610f\u3002<\/p>\n<p><strong>GitHub\uff1a<\/strong> <a href=\"https:\/\/github.com\/DeepExperience\/MMSkills\" rel=\"noopener noreferrer\">https:\/\/github.com\/DeepExperience\/MMSkills<\/a><\/p>\n<p><strong>\u7db2\u5740\uff1a<\/strong> <a href=\"https:\/\/arxiv.org\/pdf\/2605.13527\" rel=\"noopener noreferrer\">https:\/\/arxiv.org\/pdf\/2605.13527<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>MMSkills\u628a\u53ef\u91cd\u7528\u64cd\u4f5c\u7d93\u9a57\u6574\u7406\u6210\u591a\u6a21\u614b\u6280\u80fd\u5305\uff0c\u8b93\u8996\u89ba\u4ee3\u7406\u505a\u684c\u9762\u8207\u904a\u6232\u4efb\u52d9\u6642\u66f4\u7a69\u5b9a\u3002\u5b83\u4ea6\u66fe\u57282026\u5e745\u670818\u65e5\u767b\u4e0a Hugging Face Daily Papers \u7b2c1\u4f4d\u3002<\/p>\n","protected":false},"author":8,"featured_media":8335,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[116,149,197],"tags":[],"class_list":["post-8336","post","type-post","status-publish","format-standard","hentry","category-agentic","category-149","category-framework"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/8336","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=8336"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/8336\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/8335"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=8336"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=8336"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=8336"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}