diff --git a/README.md b/README.md index 6f63094..4fcf689 100644 --- a/README.md +++ b/README.md @@ -78,7 +78,7 @@ Unlike v7, v8 no longer bundles a local webgui. The web app is loaded from an on ### Nvidia Hardware acceleration users for Nvidia will need to install the container runtime provided by Nvidia on their host, instructions can be found here: -https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html +https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html We automatically add the necessary environment variable that will utilise all the features available on a GPU on the host. Once nvidia container toolkit is installed on your host you will need to re/create the docker container with the nvidia container runtime `--runtime=nvidia` and add an environment variable `-e NVIDIA_VISIBLE_DEVICES=all` (can also be set to a specific gpu's UUID, this can be discovered by running `nvidia-smi --query-gpu=gpu_name,gpu_uuid --format=csv` ). NVIDIA automatically mounts the GPU and drivers from your host into the foldingathome docker container. ## Usage