Podman Desktop (Windows / WSL2) and CUDA-GPU-Support

[brainbytes42]

Hi all,
I need to run TensorFlow with GPU-Support on a Windows-Machine.
As far as I know, since TensorFlow has abandoned Windows-GPU-Support, the only way is to use WSL and Containers. But unfortunately, I didn't manage to get it working - tried for several hours, together with a colleague... Maybe because I'm not a Linux-Expert, nor have too much experience with more complex podman usecases... We followed several tutorials from Nvidia, Microsoft and Tensorflow, but they are quite confusing and are partly inconsistent in what they say... Searched the web but nothing helpful... Feeling lost... :-(
A concise Tutorial, step by step, on how to get this to work would be really great, appreciate any help!
Thank you!