Deep Learning Frameworks on Aurora

Corey Adams, ALCF
May Dev Session Graphic featuring title and date and picture of Aurora.

Join us on May 29, 2024 at 11:00 AM for a webinar covering Deep Learning Frameworks on Aurora. 

In this developer session, Corey Adams will cover the Deep Learning Frameworks optimized by Intel for performance on Aurora.  Frameworks covered will include Tensorflow, Pytorch, and the experimental release of JAX.  Other topics include how to achieve optimal single device performance, as well as scaling up and scaling out models.  The session includes examples and demonstrations that can run on Sunspot or Aurora.

Corey J Adams is a computational scientist in the AI/ML group at ALCF, working closely with Intel on the deep learning frameworks since 2019.  His expertise includes scaling and optimizing deep learning workloads, including IO, communication, and single GPU performance.