软
Sep 14, 2026软件工程师视角：美图秀秀如何把 AIGC 塞进手机——端侧推理、GPU 渲染与性能优化实战
软件工程师视角：美图秀秀如何把 AIGC 塞进手机——端侧推理、GPU 渲染与性能优化实战 在移动端做...
Sep 14, 20261 min read0 reactions0 comments
Back to articles 
Jul 9, 2026
Tag archive
#ondeviceinference
J
Sep 19, 2026JPMorgan's 2026 Tech Trends Report Confirms What Small Models Already Proved
JPMorgan Chase dropped its 2026 Emerging Technology Trends report on March 19. According to a writeup...
Sep 19, 20267 min read0 reactions0 comments
美
Sep 18, 2026美图秀秀如何把 AIGC 塞进手机？端侧推理、跨端架构与性能优化全拆解
美图秀秀如何把 AIGC 塞进手机？端侧推理、跨端架构与性能优化全拆解 一、背景：AIGC 上手机，不只是把模型搬进去 美图秀秀里的 AI 消除、AI...
Sep 18, 20261 min read0 reactions0 comments
从
Sep 18, 2026从美图秀秀 AI 修图看工程化落地：扩散模型加速、端侧推理与实时图像渲染管线
从美图秀秀 AI 修图看工程化落地：扩散模型加速、端侧推理与实时图像渲染管线 1. 引言 美图秀秀 AI...
Sep 18, 20261 min read0 reactions0 comments
网
Sep 18, 2026网易有道翻译工程实践：大模型时代的高并发、低延迟与端侧推理架构拆解
网易有道翻译工程实践：大模型时代的高并发、低延迟与端侧推理架构拆解 ...
Sep 18, 20261 min read0 reactions0 comments
美
Sep 17, 2026美图秀秀如何把 AIGC 修图做到秒级？端侧推理、云边协同与工程化实战
美图秀秀如何把 AIGC 修图做到秒级？端侧推理、云边协同与工程化实战 当 AIGC 进入修图场景，用户对等待时间的容忍度会急剧下降。一次 AI 消除、AI 换背景、AI 超清或...
Sep 17, 20261 min read0 reactions0 comments
从
Sep 17, 2026从美图秀秀看 AIGC 应用工程化：端侧推理、实时渲染与端云协同实战
从美图秀秀看 AIGC 应用工程化：端侧推理、实时渲染与端云协同实战 一、背景：AIGC...
Sep 17, 20261 min read0 reactions0 comments
从
Sep 15, 2026从美颜到 AIGC：美图秀秀端侧推理、GPU 渲染与亿级图像架构拆解
从美颜到 AIGC：美图秀秀端侧推理、GPU 渲染与亿级图像架构拆解 从一键磨皮、滤镜到 AI 写真、文生图，美图秀秀的产品形态经历了从图像处理工具到 AIGC...
Sep 15, 20261 min read0 reactions0 comments
M
Aug 5, 2026MacPaw Teams with Liquid AI for On‑Device Inference in Its App Store
MacPaw partners with Liquid AI to bring on‑device inference to developers building for its app store, enabling a local version of the Eney assistant.
Aug 5, 20263 min read0 reactions0 comments
P
Jul 18, 2026Prism ML Introduces Bonsai 27B: A 1-bit LLM for On-Device Inference
What Changed Prism ML has unveiled Bonsai 27B, a significant advancement in on-device...
Jul 18, 20265 min read0 reactions0 comments
Best AI Gateways for On-Device and Edge Inference
The shift to AI inference at the edge and on employee devices presents unique challenges for...
Jul 9, 20267 min read0 reactions0 comments
Q
Jun 18, 2026Qwen3 in the browser, zero keys — WebLLM 0.2.83 hands-on
Run Qwen3 in Chrome via WebLLM 0.2.83 — no server needed. Step-by-step: import, stream tokens, handle cold start, pick the right model for y
Jun 18, 20267 min read0 reactions0 comments