[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-llamaweb-webgpu-llm":3,"topics-all":31,"news-related-aa53081a-e448-4087-aaa9-c822a7074bbc":50},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":24,"published_at":25,"created_at":26,"modified_at":27,"is_published":28,"publish_type":29,"image_url":13,"view_count":30},"aa53081a-e448-4087-aaa9-c822a7074bbc","LlamaWeb：WebGPU 跑 llama.cpp，16 设备吞吐 +45-69%","浏览器里跑大模型一直是个\\\"内存+跨设备兼容性\\\"双输的局面:不同浏览器、不同 GPU 上的内存占用差异巨大,加载动辄十几 GB 的模型更是家常便饭。arXiv 新论文《Llamas on the Web》(LlamaWeb)把 llama.cpp 后端重写成 WebGPU 原生实现,从根子上解决了这两个问题。LlamaWeb 的核心思路是\\\"静态化\\\":通过静态内存规划(Static Memory Planning)提前分配 WebGPU 缓冲,避免运行时反复申请释放带来的碎片;再通过可调内核库(Tunable Kernel Library)屏蔽不同设备 GPU 的差异,让同一份代码在 Apple、Intel、AMD、NVIDIA、Qualcomm 等 8 家厂商的 16 台设备上都能跑出接近原生性能。量化支持采用模板化 GPU 内核,允许扩展到任意新格式。实测数据很有说服力:相比现有浏览器 LLM 框架,内存占用降低 29-33%,在四款跨厂商 GPU 上解码吞吐提升 45-69%。在部分设备上,LlamaWeb 甚至击败了 llama.cpp 的厂商专用后端。这意味着开发者可以在 Chrome、Safari、Firefox 里直接集成 7B-13B 级别的 LLM,而不必依赖云端。对个人开发者而言,LlamaWeb 让\\\"完全本地、隐私优先\\\"的浏览器内 AI 应用有了真正的工程底座——不再因为内存墙而被迫阉割模型精度。配上 RAG 与 WebGPU 编译优化,这条\\\"零服务器 LLM\\\"的路径正在快速走向成熟。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.20706","7437aeb9-930c-4866-a2e9-48003c1a792b",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"7ac06d8e-b074-4147-abfc-ffaa4c6b8744","ai-efficiency",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"fca9258a-9430-455a-b95d-b9fae5e373a8","ai-inference",{"id":18,"name":19,"slug":19,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",{"id":21,"name":22,"slug":22,"description":13,"color":13},"b49648f9-963e-4082-8684-3d085b7358fe","quantization",[],"llamaweb-webgpu-llm","2026-06-29T18:00:00Z","2026-06-29T18:19:40.647308Z","2026-08-19T02:08:40.142862Z",true,"agent",174,[32,41],{"slug":33,"tag_slug":33,"title_zh":34,"title_en":35,"intro_zh":36,"intro_en":37,"id":38,"is_active":28,"created_at":39,"modified_at":40},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":42,"tag_slug":42,"title_zh":43,"title_en":44,"intro_zh":45,"intro_en":46,"id":47,"is_active":28,"created_at":48,"modified_at":49},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":51},[52,57,62,67,72,77],{"id":53,"title":54,"news_slug":55,"published_at":56},"645dd52f-ae6e-4931-9d74-4582feb4ecb1","Google TurboQuant：LLM推理内存压缩6倍的技术突破","google-turboquant-kv-cache-compression","2026-04-23T01:11:00+00:00",{"id":58,"title":59,"news_slug":60,"published_at":61},"89de53a6-5cee-40af-8190-1c22d628b738","23 个端侧 LLM 同台比:Artificial Analysis 把 iPhone 17 Pro 变成首个开放基准","artificial-analysis-pipette-mobile-llm-benchmark","2026-09-01T11:00:00+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"7bb4f5ec-14e0-43b6-9913-07cad82a520b","微软内部 AI 账单失控:单员工月烧 2.8 万美元,倒逼默认模型换人","microsoft-internal-ai-bill-explode-default-model-swap","2026-08-28T04:00:00+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"6fa1bc74-c98e-476b-bc4c-9ae057105ffb","ParaTempo:免训练并行推理,延迟最高降 32%、token 省三成","paratempo-temporal-confidence-parallel-reasoning","2026-08-24T17:20:00+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"ddb7bc6c-6b6e-4797-ab76-d1aeab5a3002","压缩得好≠部署得好:树莓派实测边缘 LLM,LoRA恢复模型100题押97个同答案","edge-llm-compression-raspberry-pi","2026-08-23T13:30:00+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"5da227db-f53d-4b07-a0c6-4ea16e04cd4d","CURE 用不确定性焦点做「block-parallel 投机解码」：端到端 2.66–3.49×、接受长度涨 4.2–7.5%","cure-block-parallel-speculative-decoding","2026-08-08T02:00:00+00:00"]