Tech Showcase: OpenPAI: Open Source Initiative for AI Platform in China

Open platform for AI (Open PAI) is an open-source platform for GPU cluster management and resource scheduling. PAI provides runtime environment support, GPU scheduling, and supports debugging, log collection, and port management. PAI is designed to significantly lower the operation overhead on a GPU cluster and improve the productivity of AI researchers. PAI embraces a microservices architecture: every component runs in a container and exposes an explicit service endpoint to other components. The flexibility and modular design of PAI make it an attractive platform to evaluate various research ideas.

Séries:
Microsoft Research Faculty Summit
Date:
Haut-parleurs:
Fan Yang

Taille: Microsoft Research Faculty Summit