My Pytorch Util module showed TorchVision corrupted
I am using Ubuntu 22.04. I have reinstalled pytorch in Pytorch Util multiple times. The Dental Segmentator & Total Segmentator both used the GPU well.
nvidia-smi printed out the gpu information well. Install torch in a conda environment can also access GPU.
Should I just ignore it? Thanks!
This happens to be on Linux as well, and I wasn’t able to find a solution.
If you don’t need the torchvision, you can ignore it. If you need the torchvision (e.g., Photogrammetry extension), you will have to use a different cuda version. 11.8 works for me.
There’s a discussion here: Can't install on Slicer 5.9.0 2025-05-18 on ubuntu 24.04 · Issue #34 · dalcalab/SlicerMultiverSeg · GitHub
Ultimately, installing torchvision manually with pip as suggested on the pytorch page worked.
It would be great if someone could figure out how to fix PyTorchUtils.
Thanks! /PythonSlicer -m pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu124 worked for me.
