{"id":10617,"date":"2026-08-01T03:32:51","date_gmt":"2026-07-31T19:32:51","guid":{"rendered":"https:\/\/infernews.com\/blog\/code-as-cot-for-physically-consistent-video-generation-via-an-agentic-dual-engin\/"},"modified":"2026-08-01T03:32:51","modified_gmt":"2026-07-31T19:32:51","slug":"code-as-cot-for-physically-consistent-video-generation-via-an-agentic-dual-engin","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/code-as-cot-for-physically-consistent-video-generation-via-an-agentic-dual-engin\/","title":{"rendered":"VideoCoCo \u5148\u7528\u4ee3\u78bc\u6f14\u7269\u7406\u518d\u51fa\u7247"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/08\/pasted-ef8241e5b9b7.jpg\" alt=\"cot paradigm\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e00\u6bb5\u5f71\u7247\u8981\u505a\u5230\u300c\u4f3c\u771f\u300d\u4e26\u4e0d\u53ea\u9760\u756b\u9762\u7d30\u7dfb\uff0c\u52d5\u4f5c\u56e0\u679c\u540c\u7269\u7406\u8b8a\u5316\u7ad9\u5514\u7ad9\u5f97\u4f4f\u8173\u66f4\u95dc\u9375\u3002VideoCoCo\u5c6c\u65bc<strong>\u8996\u983b\u751f\u6210\u7ba1\u7dda<\/strong>\uff0c\u8655\u7406\u7684\u6b63\u662f\u9019\u500b\u554f\u984c\uff1a\u5148\u7528 code as a chain-of-thought \u5beb\u51fa\u7269\u7406\u8349\u7a3f\uff0c\u518d\u628a\u4e2d\u6027\u767d\u6a21\u5f71\u7247\u8f49\u6210\u5beb\u5be6\u7d50\u679c\uff0c\u6e1b\u5c11\u6a21\u578b\u76f4\u63a5\u7531\u50cf\u7d20\u731c\u6e2c\u904b\u52d5\u6642\u5e38\u898b\u7684\u5931\u771f\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">VideoCoCo \u4e0d\u662f\u4e00\u6b65\u751f\u6210\u6700\u7d42\u5f71\u7247\uff0c\u800c\u662f\u5148\u7531 code agent \u7522\u751f Blender \u53ef\u57f7\u884c\u6a21\u64ec\uff0c\u8f38\u51fa\u7070\u767d\u3001\u8fd1\u4f3c clay render \u7684 proxy video\uff0c\u8b93\u5f62\u72c0\u3001\u900f\u660e\u5ea6\u3001\u8b8a\u5f62\u3001\u906e\u64cb\u8207\u904b\u52d5\u5148\u627f\u8f09\u7269\u7406\u610f\u7fa9\uff0c\u4e4b\u5f8c\u518d\u6aa2\u67e5\u9019\u6bb5\u8349\u7a3f\u662f\u5426\u7b26\u5408 physical plan\uff0c\u6700\u5f8c\u624d\u7528\u7de8\u8f2f\u6307\u4ee4\u628a proxy restyle \u6210 photorealistic video\u3002\u9019\u7a2e\u96d9\u968e\u6bb5\u6d41\u7a0b\u63db\u4f86\u8f03\u5f37\u7684\u53ef\u63a7\u6027\uff0c\u4f46\u4e5f\u4ee3\u8868\u6574\u9ad4\u93c8\u8def\u6bd4\u55ae\u6bb5\u5f0f\u751f\u6210\u66f4\u9577\uff0c\u5c0d\u4e2d\u9593\u8349\u7a3f\u54c1\u8cea\u6709\u4f9d\u8cf4\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u76ee\u524d GitHub \u5df2\u653e\u51fa\u4e94\u500b Agent Skills\u3001batch inference \u8173\u672c\u3001\u5c0d\u4e0a\u6e38 OmniWeaving \u7684 patch\uff0c\u4ee5\u53ca Hugging Face \u4e0a\u7684 tuned transformer\uff1b\u53e6\u6709 8 \u7d44 hand-checked \u7684 video-to-video triplets \u53ef\u7528\u4f86\u7406\u89e3\u8cc7\u6599\u683c\u5f0f\u8207\u8f38\u5165\u8f38\u51fa\u95dc\u4fc2\u3002README \u505c\u5728 Inference \u7ae0\u7bc0\u958b\u982d\uff0c\u672a\u898b\u5b8c\u6574\u5b89\u88dd\u8207\u57f7\u884c\u7d30\u7bc0\uff0c\u6240\u4ee5\u73fe\u968e\u6bb5\u8f03\u9069\u5408\u628a\u5b83\u8996\u70ba\u53ef\u6aa2\u67e5\u6d41\u7a0b\u8a2d\u8a08\u8207\u63a8\u7406\u7d44\u4ef6\u7684\u7814\u7a76\u578b\u9805\u76ee\uff0c\u800c\u4e0d\u662f\u5373\u88dd\u5373\u7528\u7684\u6210\u54c1\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u5148\u505a\u7269\u7406\u8349\u7a3f\uff0c\u518d\u505a\u5beb\u5be6\u5316<\/strong>\uff0c\u628a\u904b\u52d5\u56e0\u679c\u540c\u756b\u9762\u98a8\u683c\u62c6\u958b\u8655\u7406<\/li>\n<li>\u4ee5 Blender \u53ef\u57f7\u884c\u4ee3\u78bc\u627f\u8f09 process-level CoT\uff0c\u91cd\u9ede\u4e0d\u5728\u6587\u5b57\u89e3\u91cb\uff0c\u800c\u5728\u53ef\u9a57\u8b49\u7684\u6a21\u64ec\u7d50\u679c<\/li>\n<li>toy dataset \u53ea\u6709 8 \u500b\u6848\u4f8b\uff0c\u6db5\u84cb buoyancy\u3001melting\u3001surface tension\u3001boiling \u7b49\u73fe\u8c61\uff0c\u8f03\u50cf\u683c\u5f0f\u6a23\u672c<\/li>\n<li>\u5df2\u63d0\u4f9b tuned transformer\u3001inference \u8173\u672c\u8207 OmniWeaving patch\uff0c\u4f46\u516c\u958b\u8cc7\u8a0a\u672a\u8db3\u4ee5\u5b8c\u6574\u91cd\u5efa\u90e8\u7f72\u6d41\u7a0b<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u53d7\u76ca\u6700\u5927\u7684\u6703\u662f\u7814\u7a76 Agentic video generation\u3001\u591a\u6b65\u9a5f controllable generation\uff0c\u6216\u8005\u60f3\u628a\u7269\u7406\u5148\u9a57\u5e36\u5165\u8996\u983b\u6a21\u578b\u5de5\u4f5c\u6d41\u7684\u5718\u968a\u3002\u73fe\u6709\u8cc7\u6599\u672a\u898b\u5b8c\u6574\u91cf\u5316\u6307\u6a19\u6216\u5927\u898f\u6a21\u8a55\u6e2c\u7d50\u679c\uff0c\u512a\u52e2\u4e3b\u8981\u4f86\u81ea\u65b9\u6cd5\u8a2d\u8a08\u8207\u4e2d\u4ecb\u8868\u793a\u7684\u53ef\u6aa2\u67e5\u6027\uff1b\u60f3\u5224\u65b7\u751f\u6210\u7a69\u5b9a\u5ea6\u8207\u6cdb\u5316\u80fd\u529b\uff0c\u4ecd\u8981\u7b49\u66f4\u5b8c\u6574\u5be6\u9a57\u6216\u81ea\u884c\u6e2c\u8a66\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/huggingface.co\/papers\/2607.27380\" rel=\"noopener noreferrer\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/micky-li-hd\/VideoCoCo\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u60f3\u8981\u5f71\u7247\u52d5\u5f97\u5408\u7406\uff0c\u53c8\u5514\u60f3\u4e00\u958b\u59cb\u5c31\u9760\u50cf\u7d20\u751f\u6210\u78b0\u904b\u6c23\uff0cVideoCoCo\u63d0\u4f9b\u5497\u53e6\u4e00\u689d\u8def\u3002\u5b83\u5148\u628a\u7269\u7406\u904e\u7a0b\u5beb\u6210\u53ef\u57f7\u884c\u8349\u7a3f\uff0c\u518d\u4ea4\u7540\u5f71\u50cf\u6a21\u578b\u505a\u5beb\u5be6\u5316\u3002<\/p>\n","protected":false},"author":8,"featured_media":10616,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,116,31,141,199,201],"tags":[],"class_list":["post-10617","post","type-post","status-publish","format-standard","hentry","category-133","category-agentic","category-video","category-141","category-dataset-","category-skill-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10617","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=10617"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10617\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/10616"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=10617"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=10617"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=10617"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}