
{"id":10242,"date":"2026-07-16T22:57:49","date_gmt":"2026-07-16T14:57:49","guid":{"rendered":"https:\/\/infernews.com\/blog\/official-pytorch-implementation-of-quot-let-rgb-be-the-language-of-vision-quot\/"},"modified":"2026-07-16T22:57:49","modified_gmt":"2026-07-16T14:57:49","slug":"official-pytorch-implementation-of-quot-let-rgb-be-the-language-of-vision-quot","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/official-pytorch-implementation-of-quot-let-rgb-be-the-language-of-vision-quot\/","title":{"rendered":"RINO \u7528\u5716\u50cf\u7de8\u8f2f\u7d71\u4e00\u8996\u89ba\u4efb\u52d9"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/07\/teaser-8.jpg\" alt=\"RINO unifies vision under a single RGB interface: one frozen image editor, driven by a task-specific prompt, handles est\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u8207\u5176\u70ba\u6bcf\u500b\u8996\u89ba\u4efb\u52d9\u5404\u81ea\u63a5\u99c1 head\u3001decoder \u6216 adapter\uff0cRINO \u9078\u64c7\u66f4\u6fc0\u9032\u7684\u8def\u7dda\uff1a\u5168\u90e8\u6539\u5beb\u6210 RGB In, RGB Out\u3002\u5b83\u5c6c\u65bc\u4e00\u500b\u4ee5 PyTorch \u5be6\u4f5c\u7684\u7814\u7a76\u578b\u8a55\u6e2c\u8207\u5be6\u9a57\u9805\u76ee\uff0c\u6838\u5fc3\u554f\u984c\u662f\u6aa2\u67e5\u55ae\u4e00\u51cd\u7d50\u5f0f image editor\uff0c\u80fd\u5426\u540c\u6642\u8655\u7406\u8996\u89ba\u7406\u89e3\u8207\u689d\u4ef6\u751f\u6210\uff0c\u800c\u6bcb\u9808\u70ba\u6df1\u5ea6\u3001segmentation\u3001pose \u4e4b\u985e\u4efb\u52d9\u53e6\u5efa\u6a21\u7d44\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u500b\u5b9a\u4f4d\u5e36\u4f86\u7684\u5438\u5f15\u529b\u5f88\u76f4\u63a5\uff1a\u6d41\u7a0b\u7d71\u4e00\u3001\u4ecb\u9762\u7d71\u4e00\u3001\u5f8c\u7aef\u4e5f\u80fd\u4e92\u63db\u3002\u9805\u76ee\u76ee\u524d\u63a5\u4e0a\u4e09\u500b\u958b\u6e90 image-edit \u6a21\u578b\u4f5c\u70ba\u9ed1\u76d2\u5f8c\u7aef\uff0c\u5305\u62ec Qwen-Image-Edit\u3001FireRed-Image-Edit \u8207 LongCat-Image-Edit\uff1b\u4efb\u52d9\u76ee\u9304\u7d50\u69cb\u4e00\u81f4\uff0c\u6bcf\u500b task \u90fd\u6709\u7368\u7acb evaluate \u7a0b\u5f0f\u8207 output \u7d50\u679c\uff0c\u65b9\u4fbf\u9010\u9805\u8dd1 benchmark\uff0c\u6bd4\u8d77\u5404\u4efb\u52d9\u5404\u5beb\u4e00\u5957\u63a8\u7406\u908f\u8f2f\uff0c\u6574\u7406\u8207\u6bd4\u8f03\u90fd\u7701\u4e8b\u5f97\u591a\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4f46\u5b83\u7684\u53d6\u6368\u540c\u6a23\u660e\u986f\u3002RINO \u4e26\u6c92\u6709\u8a13\u7df4\u65b0\u6a21\u578b\uff0c\u4e5f\u4e0d\u505a fine-tuning\uff0c\u800c\u662f\u5805\u6301\u7528 released weights \u76f4\u63a5\u6e2c zero-shot \u8868\u73fe\uff1b\u597d\u8655\u662f\u6bd4\u8f03\u4e7e\u6de8\uff0c\u8f03\u80fd\u53cd\u6620 image editor \u672c\u8eab\u7684\u6cdb\u5316\u80fd\u529b\uff0c\u9650\u5236\u5247\u662f\u4e0a\u9650\u6703\u88ab\u539f\u751f\u7de8\u8f2f\u6a21\u578b\u7d81\u4f4f\uff0c\u5c0d\u7d50\u69cb\u5316\u8f38\u51fa\u662f\u5426\u7a69\u5b9a\u3001\u662f\u5426\u5bb9\u6613\u53d7 prompt \u8207\u6e32\u67d3\u65b9\u5f0f\u5f71\u97ff\uff0c\u4ecd\u8981\u6309\u4efb\u52d9\u9010\u500b\u770b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u91cd\u9ede\u4e0d\u662f\u8ffd\u6c42\u55ae\u4e00\u4efb\u52d9\u6700\u4f73\u6210\u7e3e\uff0c\u800c\u662f\u6e2c\u8a66\u300c\u540c\u4e00\u500b RGB \u4ecb\u9762\u300d\u80fd\u5426\u6a6b\u8de8\u591a\u985e\u8996\u89ba\u5de5\u4f5c<\/li>\n<li>\u4e09\u500b\u5f8c\u7aef\u53ef\u4e92\u63db\uff1aQwen-Image-Edit\u3001FireRed-Image-Edit\u3001LongCat-Image-Edit<\/li>\n<li>\u63a1\u7528 copied official metric code \u8a55\u5206\uff0c\u6578\u5b57\u7406\u8ad6\u4e0a\u8f03\u5bb9\u6613\u8207\u65e2\u6709\u6587\u737b\u5c0d\u9f4a<\/li>\n<li>\u90e8\u7f72\u7406\u89e3\u4e0d\u8907\u96dc\uff1a\u5b89\u88dd\u4f9d\u8cf4\u5f8c\uff0c\u6309 task \u6e96\u5099 dataset\uff0c\u518d\u9078 BACKEND \u8207\u5c0d\u61c9 MODEL \u4fbf\u53ef\u57f7\u884c\u8a55\u6e2c<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u8f03\u9069\u5408\u7559\u610f\u9019\u500b\u9805\u76ee\u7684\uff0c\u6703\u662f\u60f3\u7814\u7a76 unified vision \u4ecb\u9762\u3001\u6bd4\u8f03\u4e0d\u540c image editor \u6cdb\u5316\u529b\uff0c\u6216\u8005\u60f3\u628a\u591a\u500b benchmark \u6536\u650f\u5230\u540c\u4e00\u5de5\u4f5c\u6d41\u7684\u5718\u968a\u3002\u73fe\u6709\u8cc7\u8a0a\u672a\u5217\u51fa\u5b8c\u6574\u6210\u7e3e\u8868\uff0c\u4f46\u5b83\u5df2\u6e05\u695a\u4ea4\u4ee3\u8a55\u6e2c\u65b9\u6cd5\u3001\u8cc7\u6599\u593e\u898f\u683c\u8207\u6a21\u578b\u4f86\u6e90\uff1b\u4f5c\u70ba\u7814\u7a76\u9a57\u8b49\u5e73\u53f0\uff0c\u50f9\u503c\u5728\u65bc\u63d0\u51fa\u4e00\u5957\u53ef\u91cd\u8986\u6bd4\u8f03\u7684\u505a\u6cd5\uff0c\u800c\u4e0d\u662f\u5373\u523b\u53d6\u4ee3\u6bcf\u985e\u4efb\u52d9\u7684\u5c08\u7528\u6a21\u578b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/huggingface.co\/papers\/2607.12450\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/yangtiming\/RINO\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>GitHub<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u628a\u6df1\u5ea6\u3001\u5206\u5272\u3001\u59ff\u614b\u4f30\u8a08\u90fd\u6539\u5beb\u6210 RGB \u5716\u50cf\u8f38\u5165\u8f38\u51fa\uff0cRINO \u60f3\u9a57\u8b49\u4e00\u500b\u5927\u81bd\u65b9\u5411\uff1a\u51cd\u7d50\u5f0f image editor \u80fd\u5426\u76f4\u63a5\u517c\u4efb\u7406\u89e3\u8207\u751f\u6210\u3002<\/p>\n","protected":false},"author":8,"featured_media":10241,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,185,30,152,119,120,76,127,199],"tags":[],"class_list":["post-10242","post","type-post","status-publish","format-standard","hentry","category-133","category-qwen","category-image","category-python","category-119","category-120","category-76","category-127","category-dataset-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10242","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=10242"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10242\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/10241"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=10242"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=10242"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=10242"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}