Nvidia provides a toolkit to do this [1], getting a GPU into a container is as easy as running `podman run --device nvidia.com/gpu=all`. The process is similar

Nvidia provides a toolkit to do this [1], getting a GPU into a container is as easy as running `podman run --device nvidia.com/gpu=all`. The process is similar for Docker, but rootless Docker requires some extra steps IIRC. [1] https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/cdi-support.html https://docs.nvidia.com/datacenter/cloud-native/container-to...