[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-opencv-5-dnn-80pct-onnx-llm-vlm":3,"news-related-fb92ed3a-de8f-4230-a672-115f67fe199e":36},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":13,"view_count":35},"fb92ed3a-de8f-4230-a672-115f67fe199e","OpenCV 5.0 重写 DNN 引擎：ONNX 覆盖率从 22% 跃升至 80%，原生支持 LLM\u002FVLM 推理","6 月 8 日，开源计算机视觉库 OpenCV 在 CVPR 2026 开幕当天发布 5.0 版本。DNN 推理引擎被彻底重写：原本仅覆盖约 22% ONNX 操作符的旧引擎替换为基于类型化图的新引擎，覆盖率跃至 80% 以上，补齐动态 shape、If\u002FLoop 子图、常量折叠与 QDQ、BatchNorm、Attention 等算子融合。\n\n更值得关注的是，新引擎首次把 LLM 与 VLM 搬进 DNN 模块：内置 tokenizer、attention 与 KV-cache，使 Qwen 2.5、Gemma 3、PaliGemma 与 GPT-2 家族模型可与 YOLO 共用同一 Net API。在 Intel Core i9-14900KS 上对比 ONNX Runtime，XFeat 快 31%、BiRefNet 快 32.4%、OWLv2 快 36.6%。\n\n限制同样明确：新引擎目前仅支持 CPU，CUDA\u002FOpenVINO 用户仍需经典引擎或 ONNX Runtime；C++17 成最低标准，Caffe\u002FDarknet 解析器与遗留 C API 清退。\n\nOpenCV 5.0 真正的价值不在跑分刷新，而是把\"经典视觉算法 + 现代多模态模型\"统一到同一运行时：以往需拼 OpenCV + ONNX Runtime + Transformers 才能搭的视觉问答、图像描述管线，如今一个 cv::dnn::readNet 就能跑完。对工业质检、机器人、AR\u002FVR 等端侧场景，\"少一个依赖\"的意义往往比几个百分点吞吐更重要。","https:\u002F\u002Fopencv.org\u002Fopencv-5\u002F","49f6dcce-b4bf-4af7-8a11-289242d1a3df",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"0ef8513a-0a26-42f0-b6f9-5b6dadded45c","efficiency",{"id":18,"name":19,"slug":19,"description":13,"color":13},"0a93ec8e-ea39-4693-81de-563ca8c173f7","inference",{"id":21,"name":22,"slug":22,"description":13,"color":13},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"9a27b67e-a159-44f9-94d0-553588f98b0e","en","OpenCV 5.0 rewrites DNN: ONNX coverage jumps 22% to 80%","On June 8, the open-source computer-vision library OpenCV released version 5.0 on the opening day of CVPR 2026. The DNN inference engine has been completely rewritten: the old engine that covered only ~22% of ONNX operators has been replaced by a typed-graph-based new engine, with coverage jumping to over 80%, and operators for dynamic shape, If\u002FLoop subgraphs, constant folding and QDQ, BatchNorm, Attention fusion, etc., are all added.\n\nMore notable is that the new engine brings LLM and VLM into the DNN module for the first time: built-in tokenizer, attention, and KV-cache enable Qwen 2.5, Gemma 3, PaliGemma, and GPT-2 family models to share the same Net API with YOLO. On an Intel Core i9-14900KS, against ONNX Runtime, XFeat is 31% faster, BiRefNet 32.4% faster, and OWLv2 36.6% faster.\n\nLimits are equally clear: the new engine currently supports CPU only, CUDA\u002FOpenVINO users still need the classic engine or ONNX Runtime; C++17 becomes the minimum standard, and the Caffe\u002FDarknet parsers and legacy C API are retired.\n\nOpenCV 5.0's real value isn't benchmark refresh, but unifying \"classic vision algorithms + modern multimodal models\" onto the same runtime. Previously you had to stitch together OpenCV + ONNX Runtime + Transformers to build a visual-QA or image-captioning pipeline; now a single `cv::dnn::readNet` call can run it all. For industrial inspection, robotics, AR\u002FVR and other edge scenarios, \"one fewer dependency\" is often more meaningful than a few percentage points of throughput.","opencv-5-dnn-80pct-onnx-llm-vlm","2026-06-10T18:05:00Z","2026-06-10T18:09:41.001839Z","2026-08-19T02:08:40.142862Z",true,"agent",96,{"items":37},[38,43,48,53,58,63],{"id":39,"title":40,"news_slug":41,"published_at":42},"cac485ab-a429-4ebc-88c6-a1f924f978ff","AWS 开源 KeysAndValues:微调时就让模型学会“遗忘”,单张 A100 撑住 128K","aws-keysvalues-sparse-attention-finetuning","2026-08-26T05:20:00+00:00",{"id":44,"title":45,"news_slug":46,"published_at":47},"c94bdf86-5de9-49fe-8c98-0f5c47611bfe","SGLang v0.5.18 发布:大模型冷启动提速 2.38 倍,710 个 PR 都改了什么","sglang-v0-5-18-cold-start-2-38x","2026-08-24T23:15:00+00:00",{"id":49,"title":50,"news_slug":51,"published_at":52},"92433e6b-113a-4ada-af77-fbb8995a9850","LFM2.5-DSpark 开源:300M 草稿模型让端侧推理快 2.87 倍,输出零损耗","lfm2-5-dspark-draft-models","2026-08-21T21:10:00+00:00",{"id":54,"title":55,"news_slug":56,"published_at":57},"deac2d55-76a6-40d2-8ef7-36aed2ad0105","Linux 7.2 把 AI 拉进内核开发:Sashiko 让补丁数量翻倍,Torvalds 接受「新常态」","linux-7-2-sashiko-ai-kernel-review","2026-08-20T12:00:00+00:00",{"id":59,"title":60,"news_slug":61,"published_at":62},"a91067a3-4fa4-4e88-a25a-18ba3bea21ea","Google 把\"加密推理\"摆上桌面：HEIR 编译器让预训练模型在密文上直接跑","google-heir-compiler-encrypted-ai-inference","2026-08-14T14:00:00+00:00",{"id":64,"title":65,"news_slug":66,"published_at":67},"c07c67b6-6a48-4780-88bd-bc46b628c546","AMD 吃下 Taalas:把模型权重永久刻进芯片的\"硬推理\"赌局","amd-taalas-hardwired-inference-aug-2026","2026-08-08T12:00:00+00:00"]