Reinforcement Learning @ Hyve Catalina Virtual Workshop
Updated: Jun 29
On Friday, June 12th Hyve Solutions & Intel presented the Catalina Virtual Workshop
During this event, subject matter experts from Intel, Hyve Solutions and minds.ai presented work on Xeon usage in Deep Learning. This workshop covered Catalina, one of the industry's first 8-socket systems designed for hyperscale computing. The platform delivers a more capable, flexible infrastructure to meet the needs of today’s rapidly changing mix of deep learning (training and inferencing) plus general compute workloads, thus simplifying infrastructure planning.
From minds.ai Jeroen Bédorf presented a short overview of what reinforcement learning is, how and where it is applied, and why it has such high compute requirements. He concluded the talk with a presentation of the minds.ai DeepSim platform which can be used to perform efficient production quality reinforcement learning training.
For more information on DeepSim see: https://www.minds.ai/deepsim
For more information on the Catalina platform see: https://hyvesolutions.com/
The recording of this talk is now available online and can be watched here (17min):