Kubernetes for AI: AI-Ready Clusters with Allocatable GPU

This video shows how you can extend your #kubernetes clusters to allow Pods to claim #gpu devices and support #llms and other #ai workloads. It starts with using the #nvidia GPU Operator in #EKS, and then shows how you can take a more slimmed-down approach with the Nvidia Device Plugin in a resource-constrained #k3s cluster. Some links: - Nvidia's CUDA docs: https://developer.nvidia.com/cuda - OpenTofu repo for EKS: https://github.com/colinjlacy/tofu-eks-gpu-cluster - OpenTofu repo for K3s on Proxmox: https://github.com/colinjlacy/tofu-proxmox-gpu-ubuntu-k3s