Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Probably broken docker images

See original GitHub issue

Bug description

We probably have broken CUDA docker images - it seems that our docker images contain PyTorch version compiled with CUDA support, however torchvision without CUDA support. This is happening due to the fact we install them in separate steps and we don’t pass PyTorch wheel URL to installation step of torchvision.

It is not a high priority since torchvision is not used as often, but it suddenly made lightning-bolts testing broken.

How to reproduce the bug

No response

Error messages and logs


# Error messages and logs here please

More info

No response

cc @carmocca @akihironitta @borda

Issue Analytics

State:
Created a year ago
Reactions:1
Comments:5 (5 by maintainers)

Top GitHub Comments

1reaction

akihironittacommented, Oct 28, 2022

For the note, the current failure in https://github.com/Lightning-AI/lightning-bolts/pull/910 shouldn’t be caused by this as far as I see the logs and should be fixed by https://github.com/Lightning-AI/lightning-bolts/issues/922 and https://github.com/Lightning-AI/lightning-bolts/pull/912.

0reactions

otajcommented, Nov 7, 2022

Hi, @akihironitta, I think I agree with you, I think that I maybe grabbed some weird version of the image which installed examples (and therefore torchvision) I’m perfectly fine with having this closed