Upcoming Webinar: Deploying ML Workloads Leveraging NVIDIA Multi-Instance GPU with OpenShift and cnvrg.io
August 25, 2020 | by
While the title may pack a lot of information, tomorrow's Webinar on setting up machine learning workloads using GPUs, OpenShift and cnvrg.io should be packed with even more. Red Hat's Tushar Katarki will be joined by John Barco of NVIDIA and Yochay Ettun of cnvrg.io to explain the architecture and best practices laid out in the title.
You will learn how to:
Leverage NVIDIA MIG technology to expand the performance and value of each NVIDIA A100 Tensor Core GPU
Execute diverse workloads with MIG instances on Kubernetes infrastructure
Share GPUs and MIG instances with cnvrg.io meta-scheduler, accomplishing high utilization and data science productivity
Deliver self service attachment of MIG instances to any ML job
Build an ML pipeline with multiple jobs with right-sized GPU and guaranteed quality of service (QoS) for every job
We're hosting a new Webinar on January 28th which covers the basics of installing OpenShift and getting setup. You can register here, but the recording of this event will be online, afterwards, for ...
Next week, on December 9, Red Hat will be presenting two different talks about the usage of Red Hat OpenShift to run AI/ML workloads. From 12:30 to 1 ET, December 9, Abhinav Joshi, Red Hat's Director ...