{"id":11514,"date":"2026-08-27T03:57:27","date_gmt":"2026-08-26T19:57:27","guid":{"rendered":"https:\/\/infernews.com\/blog\/gigabrain-world-model-powered-vision-language-action-model\/"},"modified":"2026-08-27T04:02:26","modified_gmt":"2026-08-26T20:02:26","slug":"gigabrain-world-model-powered-vision-language-action-model","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/gigabrain-world-model-powered-vision-language-action-model\/","title":{"rendered":"GigaBrain-0.7 \u8b93\u6a5f\u68b0\u4eba\u8de8\u5834\u666f\u7406\u89e3\u4e26\u57f7\u884c\u4efb\u52d9"},"content":{"rendered":"\n
\"GigaBrain-0<\/figure>\n\n\n\n

\u7531\u5bb6\u5c45\u6574\u7406\u5230\u5de5\u696d\u64cd\u4f5c\uff0c\u6a5f\u68b0\u4eba\u8981\u9762\u5c0d\u7684\u4e0d\u53ea\u662f\u770b\u61c2\u756b\u9762\uff0c\u9084\u8981\u7406\u89e3\u6307\u4ee4\u3001\u9810\u6e2c\u4e0b\u4e00\u6b65\u4e26\u63a7\u5236\u4e0d\u540c\u786c\u4ef6\u3002GigaBrain-0.7 \u5c6c\u65bc Vision-language-action (VLA) embodied foundation model\uff0c\u900f\u904e\u7d71\u4e00 understanding\u3001prediction \u8207 action\uff0c\u8655\u7406\u8de8\u4efb\u52d9\u53ca\u8de8\u6a5f\u68b0\u4eba\u5f62\u614b\u7684\u6cdb\u5316\u554f\u984c\u3002<\/p>\n\n\n\n

\u9805\u76ee\u4ee5 three-system architecture \u7d44\u7e54\u6a21\u578b\u80fd\u529b\uff0c\u4e26\u628a\u9810\u8a13\u7df4\u8cc7\u6599\u64f4\u5c55\u81f3\u8d85\u904e 37,000 \u5c0f\u6642\u7684\u7570\u8cea embodied data\uff0c\u518d\u4ee5 one-stage alignment training \u540c\u6642\u512a\u5316 vision-language understanding \u548c multi-embodiment action generation\u3002\u76f8\u6bd4 GigaBrain-0 \u7cfb\u5217\u53ca\u5305\u62ec \u03c0 0.5 \u5728\u5167\u7684\u5148\u9032\u6a21\u578b\uff0c\u958b\u767c\u5718\u968a\u8072\u7a31\u5b83\u5728 foundation zero-shot capabilities\u3001language-conditioned instruction following \u53ca post-training task success rates \u5747\u6709\u660e\u986f\u63d0\u5347\u3002<\/p>\n\n\n<\/div>\n\n\n

GigaBrain-0.7 \u5df2\u63d0\u4f9b\u7a0b\u5f0f\u78bc\u3001\u6a21\u578b\u548c sample data\uff0c\u6a21\u578b\u53ca\u8cc7\u6599\u4ea6\u9023\u63a5\u81f3 Hugging Face\uff1b\u4f46 VLM evaluation code\u3001RoboColiseum\u3001RoboTwin2.0 \u548c EBench benchmark code \u4ecd\u5217\u5728\u5f85\u8fa6\u6e05\u55ae\u3002\u9019\u4ee3\u8868\u9805\u76ee\u9069\u5408\u7814\u7a76\u5718\u968a\u5148\u91cd\u73fe\u6d41\u7a0b\u53ca\u6e2c\u8a66\u8cc7\u6599\u7ba1\u7dda\uff0c\u672a\u5fc5\u5df2\u5177\u5099\u5b8c\u6574\u3001\u5373\u63d2\u5373\u7528\u7684\u6a19\u6e96\u5316\u8a55\u6e2c\u74b0\u5883\u3002<\/p>\n\n\n\n

\u8b80\u8005\u53ef\u5f9e\u4ee5\u4e0b\u5e7e\u9ede\u7406\u89e3\u5176\u50f9\u503c\u8207\u53d6\u6368\uff1a
\n– \u652f\u63f4\u591a\u7a2e\u6a5f\u68b0\u4eba\u5f62\u614b\uff0c\u76ee\u6a19\u662f\u63d0\u5347\u8de8\u786c\u4ef6\u6cdb\u5316\u80fd\u529b\u3002
\n– Maker H01 \u53ca\u4e3b\u6d41\u6a5f\u68b0\u4eba\u5e73\u53f0\u6db5\u84cb\u5bb6\u5c45\u548c\u5de5\u696d\u5834\u666f\u3002
\n– \u5927\u898f\u6a21\u7570\u8cea\u8cc7\u6599\u6709\u52a9\u64f4\u95ca\u4efb\u52d9\u7bc4\u570d\uff0c\u4f46\u4ea6\u63d0\u9ad8\u8a13\u7df4\u53ca\u786c\u4ef6\u9700\u6c42\u3002
\n– \u958b\u653e\u7a0b\u5f0f\u78bc\u3001\u6a21\u578b\u8207\u6a23\u672c\u8cc7\u6599\uff0c\u65b9\u4fbf\u7814\u7a76\u53ca\u4e8c\u6b21\u958b\u767c\u3002<\/p>\n\n\n\n

\u5c0d\u6a5f\u68b0\u4eba\u7814\u7a76\u3001\u5177\u8eab\u667a\u80fd\u53ca\u9700\u8981\u8de8\u5e73\u53f0\u63a7\u5236\u7684\u5718\u968a\uff0cGigaBrain-0.7\u63d0\u4f9b\u4e86\u8f03\u5b8c\u6574\u7684\u6a21\u578b\u3001\u8cc7\u6599\u8207\u8a13\u7df4\u5be6\u4f5c\u5165\u53e3\uff1b\u5546\u696d\u843d\u5730\u4ecd\u9700\u81ea\u884c\u9a57\u8b49\u5b89\u5168\u6027\u3001\u786c\u4ef6\u517c\u5bb9\u6027\u3001\u5ef6\u9072\u53ca\u9577\u6642\u9593\u4efb\u52d9\u7a69\u5b9a\u5ea6\u3002<\/p>\n\n\n\n

\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 GitHub<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

\u7531\u5bb6\u5c45\u6574\u7406\u5230\u5de5\u696d\u64cd\u4f5c\uff0cGigaBrain-0.7\u5617\u8a66\u4ee4\u6a5f\u68b0\u4eba\u9069\u61c9\u66f4\u591a\u8eab\u9ad4\u5f62\u614b\u8207\u5de5\u4f5c\u74b0\u5883\u3002<\/p>\n","protected":false},"author":8,"featured_media":11513,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"wpai_generated_summary":"","footnotes":""},"categories":[133,119,76,127,149,184,199,204],"tags":[],"class_list":["post-11514","post","type-post","status-publish","format-standard","hentry","category-133","category-119","category-76","category-127","category-149","category-robotic","category-dataset-","category-visionlanguageaction"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11514","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=11514"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11514\/revisions"}],"predecessor-version":[{"id":11516,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11514\/revisions\/11516"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/11513"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=11514"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=11514"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=11514"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}