Error "No NVIDIA driver on your system" about ml-workspace

[vivekruhela]

I have found an error in my command itself. I have not mentioned the --gpus argument in the docker command. My updated command is as follows:

docker run -p 8081:8080 --name "bio-dgi-workspace4" -v "${PWD}:/workspace" --gpus all --restart always mltooling/ml-workspace-gpu:latest

Now, I am not getting any such error. But the new issue I am getting is that there is no gpu utility while training my DL model, even after categorically adding the GPU_ID as an environment variable (shown below)

os.environ["CUDA_DEVICE_ORDER"] = "PCI_BUS_ID"
os.environ["CUDA_VISIBLE_DEVICES"] = "0"

Screenshot is attached below:

Can anybody tell me why there is no GPU utility?

[github-actions]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 14 days