AI computing power in Edge Industry use cases
summary
Huawei’s Atlas edge stack ships actual hardware and software: Atlas 200: half‑credit‑card, 8W inference module, Atlas 500 with Wi‑Fi/LTE and passive cooling edge station, and Atlas 300 PCIe (four Ascend chips) for server deployment. MindSpore is open source and the platform runs TensorFlow/PyTorch/NVIDIA stacks. Real deployments include Neolix delivery vehicles on public paths, manufacturing vision that raises defect detection from <90% to 99%+ and AGV fleet setups that cut robot compute power by 8–25%, extending runtime.
AI-written summary · about
chapters (9)
- 0:00Introduction and welcome
- 1:11AI evolution to edge computing
- 3:59What is edge AI computing
- 5:46AI deployment requirements in edge
- 8:42Huawei Atlas AI computing solution
- 11:52Atlas hardware devices overview
- 13:40Neolix self-driving delivery vehicles
- 16:10Manufacturing use cases in edge AI
- 20:50Q&A session
related talks
14mEdge AI: Real Solutions for Real World Problems
37mEnabling AI For Everyone - Hiai Open Platform Introduction
15mSimplifying & Scaling Edge AI
16mAI usage in telcos
13mHow Edge AI Will Save The World
05mEdge AI: Because the Cloud is Too Slow...
19mDistributed AI for Automated and Scalable Quality Inspection using EDGE Computing...
25mAI-powered digital transformation