{"id":9770,"date":"2026-06-30T02:04:26","date_gmt":"2026-06-29T18:04:26","guid":{"rendered":"https:\/\/infernews.com\/blog\/post-4d4d9800\/"},"modified":"2026-06-30T02:04:26","modified_gmt":"2026-06-29T18:04:26","slug":"post-4d4d9800","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/post-4d4d9800\/","title":{"rendered":"\u7528 vision-language-action \u628a\u4eba\u985e\u64cd\u4f5c\u8f49\u79fb\u5230\u6a5f\u68b0\u4eba\u7684\u96d9\u624b"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/06\/pasted-8dc1556086ea.jpg\" alt=\"Hugging Face\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u4eba\u985e\u64cd\u4f5c\u6578\u64da\u5ec9\u50f9\u3001\u8c50\u5bcc\u4e14\u591a\u6a23\u5316\uff0c\u4f7f\u5176\u6210\u70ba\u64f4\u5c55\u6a5f\u5668\u4eba\u5b78\u7fd2\u898f\u7bc4\u6700\u6709\u524d\u666f\u7684\u8cc7\u6e90\u4e4b\u4e00\u3002\u7136\u800c\uff0c\u5c07\u4eba\u985e\u6280\u80fd\u9077\u79fb\u5230\u6a5f\u5668\u4eba\u4ecd\u7136\u56f0\u96e3\u91cd\u91cd\uff1a\u4ee5\u5f80\u7684\u5927\u591a\u6578\u5de5\u4f5c\u90fd\u5c07\u4eba\u985e\u8996\u70ba\u53e6\u4e00\u7a2e\u96d9\u624b6\u81ea\u7531\u5ea6\uff086DoF\uff09\u7684\u5177\u8eab\u6a21\u578b\uff0c\u9019\u5b58\u5728\u5169\u500b\u554f\u984c\uff1a\u624b\u90e8\u59ff\u614b\u4f30\u8a08\u5b58\u5728\u566a\u8072\uff0c\u4e26\u4e14\u4eba\u985e\u624b\u6307\u7684\u63a5\u89f8\u6a21\u5f0f\u8207\u4e26\u806f\u6a5f\u68b0\u81c2\u7684\u63a5\u89f8\u6a21\u5f0f\u5b58\u5728\u6839\u672c\u5dee\u7570\uff0c\u5c0e\u81f4\u624b\u8155\u65cb\u8f49\u8207\u6a5f\u68b0\u81c2\u64cd\u4f5c\u5728\u8a9e\u7fa9\u4e0a\u4e0d\u4e00\u81f4\u3002\u6211\u5011\u8a8d\u70ba\uff0c\u5f9e\u4eba\u985e\u8cc7\u6599\u4e2d\u5b78\u7fd2\u5305\u542b\u65cb\u8f49\u7684\u52d5\u4f5c\u8a0a\u865f\u4e26\u975e\u6700\u512a\u65b9\u6848\uff0c\u56e0\u6b64\u63d0\u51fa\u4e86\u4e00\u7a2e \u6a4b\u63a5\u52d5\u4f5c\u8868\u793a\uff1a\u5728\u521d\u59cb\u982d\u90e8-\u651d\u5f71\u6a5f\u5ea7\u6a19\u7cfb\u5167\u7684\u76f8\u5c0d\u624b\u8155\u5e73\u79fb\uff0c\u9019\u662f\u4e00\u500b\u4eba\u985e\u548c\u6a5f\u5668\u4eba\u5171\u4eab\u7684\u52d5\u4f5c\u7a7a\u9593\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5982\u4f55\u628a\u4eba\u985e\u64cd\u4f5c\u8cc7\u6599\u8f49\u79fb\u5230\u4f7f\u7528 parallel grippers \u7684\u96d9\u624b\u6a5f\u68b0\u4eba\uff0c\u91cd\u9ede\u65b9\u6cd5\u662f\u7528\u300crelative wrist translation within the initial head-camera frame\u300d\u4f5c\u70ba bridging action representation\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4f5c\u8005\u8a8d\u70ba\u628a\u4eba\u985e\u76f4\u63a5\u7576\u6210\u53e6\u4e00\u7a2e bi-manual 6DoF embodiment \u4e26\u4e0d\u7406\u60f3\uff0c\u56e0\u70ba\u624b\u90e8\u59ff\u614b\u4f30\u8a08\u672c\u8eab\u6709\u566a\u8072\uff0c\u800c\u4e14\u4eba\u624b\u624b\u6307\u63a5\u89f8\u6a21\u5f0f\u8207 parallel gripper \u6709\u672c\u8cea\u5dee\u7570\u3002\u8207\u5176\u786c\u5b78\u5305\u542b\u65cb\u8f49\u7684\u52d5\u4f5c\u8a0a\u865f\uff0c\u9019\u9805\u5de5\u4f5c\u6539\u70ba\u53ea\u4fdd\u7559\u66f4\u5bb9\u6613\u8de8\u4eba\u985e\u8207\u6a5f\u68b0\u4eba\u5171\u4eab\u7684\u5e73\u79fb\u8cc7\u8a0a\uff0c\u6e1b\u5c11 embodiment mismatch\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4f5c\u8005\u5efa\u7acb\u4e86\u4e00\u500b \u03c00-like vision-language-action model\uff0c\u914d\u5408 interleaved action tokens \u8207 attention masking\uff0c\u8655\u7406\u4e0d\u540c embodiment \u53ef\u80fd\u7f3a\u5c11\u67d0\u4e9b\u52d5\u4f5c\u6210\u5206\u7684\u554f\u984c\u3002\u9019\u7a2e\u8a2d\u8a08\u7684\u610f\u7fa9\uff0c\u5728\u65bc\u6a21\u578b\u4e0d\u9700\u8981\u5047\u8a2d\u4eba\u8207\u6a5f\u68b0\u4eba\u64c1\u6709\u5b8c\u5168\u76f8\u540c\u7684\u63a7\u5236\u7dad\u5ea6\uff0c\u8f03\u9069\u5408\u8de8\u8f09\u9ad4\u6280\u80fd\u9077\u79fb\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u4ee5 wrist translation \u53d6\u4ee3\u5b8c\u6574 6DoF human actions\uff0c\u964d\u4f4e\u4eba\u624b\u5230\u593e\u722a\u7684\u8868\u793a\u843d\u5dee<\/li>\n<li>\u63a1\u7528 vision-language-action \u67b6\u69cb\uff0c\u4e26\u52a0\u5165 interleaved action tokens \u8207 attention masking<\/li>\n<li>\u5728 novel bi-manual manipulation tasks \u4e0a\uff0c\u8f03 noisy 6DoF human actions \u6709\u66f4\u6709\u6548\u7684\u77e5\u8b58\u8f49\u79fb<\/li>\n<li>\u6548\u679c\u6703\u96a8 human data \u6578\u91cf\u589e\u52a0\u800c\u63d0\u5347\uff0c\u8aaa\u660e\u65b9\u6cd5\u5177\u5099\u4e00\u5b9a\u64f4\u5c55\u6027<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u9805\u5167\u5bb9\u8f03\u63a5\u8fd1\u65b9\u6cd5\u8ad6\u8207\u8868\u793a\u5b78\u7fd2\u5206\u6790\uff0c\u800c\u4e0d\u662f\u90e8\u7f72\u6307\u5357\u3002\u9801\u9762\u6c92\u6709\u5217\u51fa\u63a8\u8ad6\u6846\u67b6\u3001\u786c\u9ad4\u9700\u6c42\u3001v2 \u6a94\u6848\u66f4\u65b0\u3001chat template \u6216 MTP draft speculation \u7b49\u8cc7\u8a0a\uff1b\u80fd\u78ba\u5b9a\u7684\u662f\uff0c\u5b83\u91dd\u5c0d Robotic \u6280\u80fd\u8f49\u79fb\u63d0\u51fa\u4e00\u7a2e\u66f4\u8cbc\u8fd1\u593e\u722a\u6a5f\u68b0\u4eba\u63a7\u5236\u9700\u6c42\u7684\u52d5\u4f5c\u62bd\u8c61\uff0c\u9069\u5408\u95dc\u6ce8 imitation learning\u3001cross-embodiment transfer \u8207\u96d9\u624b\u64cd\u4f5c\u7814\u7a76\u7684\u4eba\u95b1\u8b80\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/huggingface.co\/papers\/2606.28133\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/arxiv.org\/pdf\/2606.28133\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>Paper<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u9019\u9805\u8ad6\u6587\u63d0\u51fa\u4e00\u7a2e\u6a4b\u63a5\u52d5\u4f5c\u8868\u793a\uff0c\u6539\u5584\u4eba\u985e\u793a\u7bc4\u5411\u96d9\u624b\u593e\u722a\u6a5f\u68b0\u4eba\u7684\u6280\u80fd\u8f49\u79fb\u3002\u91cd\u9ede\u5728\u65bc\u907f\u958b\u5608\u96dc\u76846DoF\u624b\u90e8\u59ff\u614b\u3002<\/p>\n","protected":false},"author":8,"featured_media":9769,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[169,176,119,149,184,201],"tags":[],"class_list":["post-9770","post","type-post","status-publish","format-standard","hentry","category-169","category-176","category-119","category-149","category-robotic","category-skill-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9770","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=9770"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9770\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/9769"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=9770"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=9770"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=9770"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}