This guide will help people that wants to use their nvidia gpu

On (proxmox) host

• Download latest nvidia drivers: https://www.nvidia.com/en-us/drivers/unix/
• For this guide, it will be https://us.download.nvidia.com/XFree86/Linux-x86_64/580.95.05/NVIDIA-Linux-x86_64-580.95.05.run
• make it executable: chmod +x NVIDIA-Linux-x86_64-580.95.05.run
• Install proxmox headers: apt install pve-headers
• Install nvidia driver: ./NVIDIA-Linux-x86_64-580.95.05.run --dkms
• Reboot
• Then check if it worked: nvidia-smi

In LXC container

• Create your container
• Make device passthrough (using lxc config or in proxmox ui) for:
  • dev0: /dev/nvidia0
  • dev1: /dev/nvidiactl
  • dev2: /dev/nvidia-uvm
  • dev3: /dev/nvidia-uvm-tools
• Download and install nvidia driver, but this time with --no-kernel-module param instead of --dkms
• Check if it works with nvidia-smi

In LXC container: Add nvidia runtime to docker

• Follow https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html