No module named 'torch' after 4.1 upgrade

[yodamoppet]

This fix appears to work for us as well on a CentOS system.

Reference:

3DFlex Dependencies; Building pycuda

Hey folks, Had the same issue and figured out how to fix it. I think @rbs_sci is right and it’s a mix of cuda version which is the issue. I simply change the line of cryosparcw from conda install -y cuda-nvcc=11.7 cuda-toolkit=11.7 -c nvidia to conda install -y cuda-nvcc=11.7 cuda-toolkit=11.7 -c “nvidia/label/cuda-11.7.0” This force the usage of 11.7 for all packages and it seems to work afterwards Hope that helps.

I did have to first revert the system:

./bin/cryosparcw forcedeps

Then edit cryosparcw and change from:

conda install -y cuda-nvcc=11.7 cuda-toolkit=11.7 -c nvidia

To:

conda install -y cuda-nvcc=11.7 cuda-toolkit=11.7 -c nvidia/label/cuda-11.7.0

Then finally run the 3dflex installer:

./bin/cryosparcw install-3dflex

No more lengthy errors, and users report that jobs appear ok so far.