{"id":7176,"date":"2025-07-29T05:11:00","date_gmt":"2025-07-28T20:11:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7176"},"modified":"2025-07-27T10:16:41","modified_gmt":"2025-07-27T01:16:41","slug":"docopilot-improving-multimodal-models-for-document-level-understanding","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7176","title":{"rendered":"Docopilot: Improving Multimodal Models for Document-Level Understanding\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Docopilot: Improving Multimodal Models for Document-Level Understanding\u00a0<\/strong>[87.6]<br>\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u6587\u66f8\u306e\u8a73\u7d30\u306a\u7406\u89e3\u3092\u652f\u63f4\u3059\u308b\u305f\u3081\u306b,\u9ad8\u54c1\u8cea\u306a\u6587\u66f8\u30ec\u30d9\u30eb\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8Doc-750K\u3092\u63d0\u6848\u3059\u308b\u3002 \u3053\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u306f\u3001\u3055\u307e\u3056\u307e\u306a\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u69cb\u9020\u3001\u5e83\u7bc4\u306a\u30af\u30ed\u30b9\u30da\u30fc\u30b8\u4f9d\u5b58\u95a2\u4fc2\u3001\u304a\u3088\u3073\u5143\u306e\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u304b\u3089\u6d3e\u751f\u3057\u305f\u5b9f\u969b\u306e\u8cea\u554f\u3068\u56de\u7b54\u306e\u30da\u30a2\u304c\u542b\u307e\u308c\u3066\u3044\u308b\u3002 \u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u57fa\u3065\u3044\u3066\u3001RAG\u306b\u983c\u308b\u3053\u3068\u306a\u304f\u3001\u6587\u66f8\u30ec\u30d9\u30eb\u306e\u4f9d\u5b58\u95a2\u4fc2\u3092\u6b63\u78ba\u306b\u51e6\u7406\u3067\u304d\u308b\u30cd\u30a4\u30c6\u30a3\u30d6\u306a\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u30e2\u30c7\u30eb\u3067\u3042\u308bDocopilot\u3092\u958b\u767a\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2507.14675v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2507.14675v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Sat, 19 Jul 2025 16:03:34 GMT)<\/li>\n\n\n\n<li>\u5927\u898f\u6a21\u306a\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30ebDocumentUnderstanding\u7528\u30c7\u30fc\u30bf\u306e\u69cb\u7bc9\u3068InternVL2\u30d9\u30fc\u30b9\u306e\u30e2\u30c7\u30eb\u69cb\u7bc9\u3002\u300cThe proposed Docopilot-8B shows a notable improvement over baseline models [73], achieving a +19.9% accuracy gain compared to InternVL2-8B and surpassing InternVL2-26B with less than 31% of the inference latency. Additionally, Docopilot-2B uses fewer parameters (less than 10%) while exhibiting comparable performance to the 10\u00d7 larger InternVL2-26B.\u300d\u3068\u6027\u80fd\u5411\u4e0a\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/OpenGVLab\/Docopilot\">OpenGVLab\/Docopilot: [CVPR 2025] Docopilot: Improving Multimodal Models for Document-Level Understanding<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[117,251],"class_list":["post-7176","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-document-understanding","tag-mllm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7176","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7176"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7176\/revisions"}],"predecessor-version":[{"id":7177,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7176\/revisions\/7177"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7176"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7176"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7176"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}