Explain whether the instructions on NVIDIA's site to install the toolkit are a prerequisite, an option, a corequisite, or an alternative to the subsequent instructions on the page, and/or whether or not the instructions are mutually exclusive.
I have spent well over 40 hours troubleshooting CUDA in containers on WSL2 related to these instructions. Please disambiguate these instructions for us all.
Documentation Issue
https://learn.microsoft.com/en-us/windows/wsl/tutorials/gpu-compute#setting-up-nvidia-cuda-with-docker step 3 reads:
Clearly, it is not clear whether the following steps are meant to follow the steps you would follow if you follow the link!
Link to documentation page
https://learn.microsoft.com/en-us/windows/wsl/tutorials/gpu-compute#setting-up-nvidia-cuda-with-docker
Suggested Improvements
Explain whether the instructions on NVIDIA's site to install the toolkit are a prerequisite, an option, a corequisite, or an alternative to the subsequent instructions on the page, and/or whether or not the instructions are mutually exclusive.