Hello! I am trying to use pytorch for the first time in a while and am facing some problems regarding versioning. I have been trying to follow installation instructions from a specific github repository relying on pytorch ( ```
conda install pytorch==1.11.0 cudatoolkit=11.3 -c pytorch -y
conda install pyg::pytorch-scatter=2.0.9 -y
I'm currently working using Ubuntu on WSL and conda, and running nvidia-smi returns the following:
So how should I approach this? Everything I've done so far resulted in my GPU not being recognized by torch (I've tried to install cuda toolkit 11.3 and 11.8, should I try more recent versions?).
Thanks!