nvidia-container-toolkit

The NVIDIA Container Toolkit allows users to build and run GPU accelerated containers.

Download this Image

The image is available on cgr.dev:

docker pull cgr.dev/chainguard/nvidia-container-toolkit:latest

Usage

helm repo add nvidia https://helm.ngc.nvidia.com/nvidia
helm upgrade --install gpu-operator nvidia/gpu-operator \
    -n gpu-operator \
    --create-namespace \
    --set toolkit.repository=cgr.dev/chainguard \
    --set toolkit.image=nvidia-container-toolkit \
    --set toolkit.version=latest

• Refer to values.yaml file for more configuration options.

You'll want to make sure the gpu-operator chart is up-to-date and use the latest operator tag that's within the compatibility matrix.

You need GPU nodes to run the operator as it will schedule Deployments and DaemonSets on nodes with GPUs.

If you want to learn more about how we are testing this image, please refer to the TESTING.md file.