Kubernetes for AI: Pass-Through GPU in a Linux Machine

First video in a series about deploying #ai models on #kubernetes. This one covers how to set up a host machine to pass an #nvidia #gpu device through to a hypervisor, so that it can be claimed by a #virtualmachine. This is a necessary step to deploying a GPU node in your Kubernetes cluster if you're managing the hardware and private cloud that your cluster will run on. Come back for the next video, where I'll show how I used #opentofu (or #terraform) and the #proxmox provider to stand up VMs in Proxmox, and then wired them together with a #k3s distribution of Kubernetes, including a GPU node sitting alongside the regular worker nodes.