{"id":5626,"date":"2024-03-16T12:04:35","date_gmt":"2024-03-16T04:04:35","guid":{"rendered":"https:\/\/www.1ai.net\/?p=5626"},"modified":"2024-03-16T12:04:35","modified_gmt":"2024-03-16T04:04:35","slug":"ai%e5%ae%9e%e6%97%b6%e7%bb%98%e7%94%bb%e7%b3%bb%e7%bb%9fstreammultidiffusion-%e6%94%af%e6%8c%81%e5%b1%80%e9%83%a8%e6%b6%82%e6%8a%b9%e6%8f%90%e7%a4%ba%e7%94%9f%e6%88%90%e5%9b%be%e7%89%87","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/5626.html","title":{"rendered":"AI real-time painting system StreamMultiDiffusion supports local painting + prompts to generate pictures"},"content":{"rendered":"<p>\u8fd1\u671f\uff0c\u4e00\u7bc7\u540d\u4e3a&#8221;<a href=\"https:\/\/www.1ai.net\/en\/tag\/streammultidiffusion\" title=\"\u3010\u67e5\u770b\u542b\u6709[StreamMultiDiffusion]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">StreamMultiDiffusion<\/a>&#8220;\u7684\u8bba\u6587\u63d0\u51fa\u4e86\u4e00\u79cd\u65b0\u9896\u7684\u5b9e\u65f6\u3001\u4ea4\u4e92\u5f0f\u7684\u6587\u672c\u5230\u56fe\u50cf\u751f\u6210\u7cfb\u7edf\u3002\u8fd9\u79cd\u7cfb\u7edf\u80fd\u591f\u6839\u636e\u7528\u6237\u63d0\u4f9b\u7684\u624b\u7ed8\u533a\u57df\u548c\u76f8\u5e94\u7684\u8bed\u4e49\u6587\u672c\u63d0\u793a\u6765\u751f\u6210\u56fe\u50cf\uff0c\u4e3a\u4e13\u4e1a\u56fe\u50cf\u521b\u4f5c\u8005\u63d0\u4f9b\u4e86\u4e00\u4e2a\u5f3a\u5927\u7684\u5de5\u5177\uff0c\u53ef\u4ee5\u7528\u4e8e\u5feb\u901f\u539f\u578b\u8bbe\u8ba1\u548c\u521b\u610f\u63a2\u7d22\u3002<\/p>\n<p class=\"article-content__img\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5627\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/03\/6384611163611304142720431.png\" alt=\"\" width=\"742\" height=\"480\" \/><\/p>\n<p>\u9879\u76ee\u5730\u5740:https:\/\/github.com\/ironjr\/StreamMultiDiffusion<\/p>\n<p>\u6269\u6563\u6a21\u578b\u5728\u6587\u672c\u5230\u56fe\u50cf\u5408\u6210\u9886\u57df\u53d6\u5f97\u4e86\u5de8\u5927\u6210\u529f\uff0c\u6210\u4e3a\u4e86\u56fe\u50cf\u751f\u6210\u548c\u7f16\u8f91\u7684\u6709\u524d\u9014\u7684\u5019\u9009\u8005\u3002\u7136\u800c\uff0c\u5c06\u8fd9\u4e9b\u6a21\u578b\u7528\u4e8e\u5b9e\u9645\u5e94\u7528\u4ecd\u9762\u4e34\u4e24\u5927\u6311\u6218:\u4e00\u662f\u9700\u8981\u66f4\u5feb\u7684\u63a8\u7406\u901f\u5ea6\uff0c\u4e8c\u662f\u9700\u8981\u66f4\u667a\u80fd\u7684\u6a21\u578b\u63a7\u5236\u3002\u8fd9\u4e24\u4e2a\u76ee\u6807\u9700\u8981\u540c\u65f6\u6ee1\u8db3\uff0c\u624d\u80fd\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u53d1\u6325\u4f5c\u7528\u3002\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e9b\u6311\u6218\uff0c\u4f5c\u8005\u63d0\u51fa\u4e86StreamMultiDiffusion\u6846\u67b6\u3002<\/p>\n<p>\u8be5\u6846\u67b6\u662f<span class=\"spamTxt\">\u7b2c\u4e00<\/span>\u4e2a\u5b9e\u65f6\u57fa\u4e8e\u533a\u57df\u7684\u6587\u672c\u5230\u56fe\u50cf\u751f\u6210\u6846\u67b6\u3002\u901a\u8fc7\u7a33\u5b9a\u5feb\u901f\u63a8\u7406\u6280\u672f\u5e76\u91cd\u6784\u6a21\u578b\u4e3a\u65b0\u63d0\u51fa\u7684\u591a\u63d0\u793a\u6d41\u6279\u5904\u7406\u67b6\u6784\uff0c\u5b9e\u73b0\u4e86\u6bd4\u73b0\u6709\u89e3\u51b3\u65b9\u6848\u66f4\u5feb\u7684\u5168\u666f\u56fe\u751f\u6210\u901f\u5ea6\uff0c\u5e76\u5728\u5355\u4e2aRTX2080Ti GPU\u4e0a\u5b9e\u73b0\u4e86\u57fa\u4e8e\u533a\u57df\u7684\u6587\u672c\u5230\u56fe\u50cf\u5408\u6210\u76841.57FPS\u751f\u6210\u901f\u5ea6\u3002<\/p>\n<p>\u8be5\u6846\u67b6\u5f15\u5165\u4e86\u51e0\u79cd\u5173\u952e\u6280\u672f\u3002\u9996\u5148\u662fLatent Pre-Averaging\uff0c\u5728\u63a8\u7406\u7684\u6bcf\u4e2a\u6b65\u9aa4\u4e2d\uff0c\u5148\u5bf9\u4e2d\u95f4\u6f5c\u5728\u8868\u793a\u8fdb\u884c\u5e73\u5747\uff0c\u4ee5\u9002\u5e94\u5feb\u901f\u63a8\u7406\u7b97\u6cd5\u3002\u5176\u6b21\u662fMask-Centering Bootstrapping\uff0c\u5728\u751f\u6210\u8fc7\u7a0b\u7684\u524d\u51e0\u6b65\u4e2d\uff0c\u5c06\u6bcf\u4e2a\u906e\u7f69\u7684\u4e2d\u5fc3\u70b9\u5bf9\u9f50\u5230\u56fe\u50cf\u4e2d\u5fc3\uff0c\u4ee5\u786e\u4fdd\u5bf9\u8c61\u4e0d\u4f1a\u88ab\u906e\u7f69\u8fb9\u7f18\u5207\u65ad\u3002\u518d\u6b21\u662fQuantized Masks\uff0c\u901a\u8fc7\u91cf\u5316\u906e\u7f69\u6765\u63a7\u5236\u63d0\u793a\u906e\u7f69\u7684\u7d27\u5bc6\u5ea6\uff0c\u4ece\u800c\u5728\u4e0d\u540c\u566a\u58f0\u6c34\u5e73\u4e0b\u5e73\u6ed1\u5730\u878d\u5408\u751f\u6210\u533a\u57df\u3002<\/p>\n<p>\u6b64\u5916\uff0cStreamMultiDiffusion\u8fd8\u5f15\u5165\u4e86\u4e00\u4e2a\u540d\u4e3aSemantic Palette\u7684\u65b0\u6982\u5ff5\uff0c\u8fd9\u662f\u4e00\u79cd\u4ea4\u4e92\u5f0f\u56fe\u50cf\u751f\u6210\u8303\u5f0f\uff0c\u5141\u8bb8\u7528\u6237\u901a\u8fc7\u624b\u7ed8\u533a\u57df\u548c\u6587\u672c\u63d0\u793a\u5b9e\u65f6\u751f\u6210\u9ad8\u8d28\u91cf\u56fe\u50cf\u3002\u8fd9\u79cd\u65b9\u6cd5\u7c7b\u4f3c\u4e8e\u4f7f\u7528\u753b\u7b14\u5728\u753b\u5e03\u4e0a\u7ed8\u5236\uff0c\u4f46\u4f7f\u7528\u7684\u662f\u6587\u672c\u63d0\u793a\u548c\u906e\u7f69\u3002\u4f8b\u5982\uff0c\u7528\u6237\u53ef\u4ee5\u5728\u7ea2\u8272\u533a\u57df\u751f\u6210\u4eba\u7269\uff0c\u5728\u8033\u6735\u548c\u5c3e\u5df4\u533a\u57df\u6807\u8bb0\u4e3a\u72d7\uff0c\u7cfb\u7edf\u4f1a\u6839\u636e\u6d82\u62b9\u7684\u533a\u57df\u751f\u6210\u957f\u7740\u72d7\u72d7\u8033\u6735\u548c\u5c3e\u5df4\u7684\u4eba\u7269\u3002<\/p>\n<p>\u8bba\u6587\u4e2d\u7684\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0cStreamMultiDiffusion\u5728\u5168\u666f\u56fe\u751f\u6210\u548c\u57fa\u4e8e\u533a\u57df\u7684\u6587\u672c\u5230\u56fe\u50cf\u5408\u6210\u65b9\u9762\uff0c\u76f8\u6bd4\u4e8e\u73b0\u6709\u7684MultiDiffusion\u65b9\u6cd5\uff0c\u5b9e\u73b0\u4e86\u663e\u8457\u7684\u901f\u5ea6\u63d0\u5347\uff0c\u540c\u65f6\u4fdd\u6301\u4e86\u56fe\u50cf\u8d28\u91cf\u3002\u8fd9\u8bc1\u660e\u4e86\u8be5\u7cfb\u7edf\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u7684\u5de8\u5927\u6f5c\u529b\u548c\u4ef7\u503c\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Recently, a paper entitled \u201cStreamMultiDiffusion\u201d presented an innovative, real-time, interactive text-to-image system. Such a system can generate images based on hand-drawn areas and corresponding semantic text tips provided by users and provides professional image creators with a powerful tool for rapid prototype design and creative exploration. Project address: https:\/\/github.com\/ironjr\/StreamMultiDiffusion proliferation model has achieved great success in the field of text-to-image synthesis, becoming a promising candidate for image generation and editing. However, the application of these models to operational applications still faces two major challenges: the need for faster<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[146],"tags":[1715,1716],"collection":[],"class_list":["post-5626","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-streammultidiffusion"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5626","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=5626"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5626\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=5626"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=5626"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=5626"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=5626"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}