从美图秀秀看端侧 AI 图像工程:实时美颜、模型压缩与移动端性能优化实战
从美图秀秀看端侧 AI 图像工程:实时美颜、模型压缩与移动端性能优化实战 1. 背景:端侧 AI 图像处理的挑战 移动端图像应用正在从“滤镜工具”走向“实时...
Tag archive
从美图秀秀看端侧 AI 图像工程:实时美颜、模型压缩与移动端性能优化实战 1. 背景:端侧 AI 图像处理的挑战 移动端图像应用正在从“滤镜工具”走向“实时...
Multiverse Computing says an IBM Heron helped generate a small synthetic-data component for healing a compressed 438B classical model, without showing that quantum hardware caused its reported gains.
拆解美图秀秀端侧 AIGC:模型压缩、GPU 推理与亿级图片架构实践 ...
千万 DAU 下的端侧 AIGC:美图秀秀如何把图像大模型塞进手机并跑出极致性能? 一、问题背景:千万 DAU 与端侧 AIGC 的碰撞 美图秀秀作为千万...
Here's the sentence that should stop you mid-scroll: a compressed, 4-bit model that outperforms its...
Canonical version:...
Nvidia's Nemotron-3-Puzzle compresses a 120-billion-parameter model down to 75 billion using a multi-stage pipeline of pruning, distillation, and quantization -- roughly doubling throughput on a B200 node and lifting single-H100 concurrency at long c

Deploying machine learning models on edge devices, such as smartphones, IoT sensors, embedded...
The 300MB Model That Runs in 40ms BERT-base sits at 110M parameters and 440MB on disk....
The pervasive adoption of large language models (LLMs) and other deep neural networks has ushered in...