How to use Charmed Kubeflow with NVIDIA GPUs

If you want to use Kubeflow backed with NVIDIA GPUs, your nodes must have runtimeClassName: nvidia parameter specified.

Contents:

MicroK8s

When deploying with MicroK8s , this parameter is added by default when enabling the gpu addon.

microk8s enable gpu

Other Kubernetes clusters

In case of other Kubernetes deployments, this parameter may not be specified. Please refer to this documentation for a general guide on how to setup NVIDIA GPU Operator.