Amazon EKS cluster configuration for AI/ML workloads

This section is designed to help you configure Amazon EKS clusters optimized for AI/ML workloads. You’ll find guidance on running GPU-accelerated containers using Linux and Windows optimized AMIs, setting up training clusters with Elastic Fabric Adapter (EFA) for high-performance networking, and creating inference clusters with AWS Inferentia instances, including prerequisites, step-by-step procedures, and deployment considerations.

Topics

• Run GPU-accelerated containers (Linux on EC2)

• Run GPU-accelerated containers (Windows on EC2 G-Series)

• Run machine learning training on Amazon EKS with Elastic Fabric Adapter

• Use AWS Inferentia instances with Amazon EKS for Machine Learning