Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update previous-versions.md #1813

Open
wants to merge 25 commits into
base: site
Choose a base branch
from
Open

Conversation

Tony-Y
Copy link

@Tony-Y Tony-Y commented Nov 11, 2024

Use CUDA 11.1 instead of CUDA 11.3 for PyTorch 1.8.1, 1.9.0, and 1.9.1.

This patch will close three issues: #1403, #1033, and #925.

This work comes from: https://discuss.pytorch.org/t/cant-find-package-of-pytorch1-8-1-and-cudatoolkit-11-3/212679

Use CUDA 11.1 instead of CUDA 11.3 for PyTorch 1.8.1, 1.9.0, and 1.9.1.
Copy link

linux-foundation-easycla bot commented Nov 11, 2024

CLA Signed

The committers listed above are authorized under a signed CLA.

Copy link

netlify bot commented Nov 11, 2024

Deploy Preview for pytorch-dot-org-preview ready!

Name Link
🔨 Latest commit 6ac1cd5
🔍 Latest deploy log https://app.netlify.com/sites/pytorch-dot-org-preview/deploys/676b747da99cce00088fc6e9
😎 Deploy Preview https://deploy-preview-1813--pytorch-dot-org-preview.netlify.app
📱 Preview on mobile
Toggle QR Code...

QR Code

Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify site configuration.

@Tony-Y
Copy link
Author

Tony-Y commented Dec 17, 2024

I have checked PyTorch installations for CUDA 11.1 and 11.3 using GitHub Actions.

CUDA 11.1: https://github.com/Tony-Y/check-conda-pytorch/actions/runs/12367435889
The CUDA version of PyTroch was installed for CUDA 11.1.
cuda-111

CUDA 11.3: https://github.com/Tony-Y/check-conda-pytorch/actions/runs/12367435900
The CUDA version of PyTroch was NOT installed for CUDA 11.3. Instead, the CPU version was installed.
cuda-113

Edit
OS: Ubuntu 22.04, Windows Server 2022
Python: 3.9
PyTorch: 1.8.1, 1.9.0, 1.9.1
CUDA Toolkit: 11.1, 11.3

@Tony-Y
Copy link
Author

Tony-Y commented Dec 24, 2024

I have checked if it is possible to import torchvision and torchaudio. On Windows, we have to install pillow and soundfile via PyPI. The CUDA version of pillow emits an error:

ImportError: DLL load failed while importing _imaging: The specified module could not be found.

because jpeg-9b is installed by the requirements for torchvision. soundfile is required for enabling the SoundFile backend in torchaudio. Refer to https://pytorch.org/audio/0.8.0/backend.html.

On Windows, Torchaudio 0.8.1 issues a warning:

UserWarning: torchaudio C++ extension is not available.

because

"sox" and "sox_io" backends require C++ extension module, which is included in Linux/macOS binary distributions.
These backends are not available on Windows.

CUDA 11.1: https://github.com/Tony-Y/check-conda-pytorch/actions/runs/12468852141/job/34800857037
The CUDA version of PyTroch and Torchvision were installed for CUDA 11.1.
windows-cu111

CUDA 11.3: https://github.com/Tony-Y/check-conda-pytorch/actions/runs/12468852141/job/34800857289
The CUDA version of PyTroch and Torchvision were NOT installed for CUDA 11.3. Instead, the CPU version was installed.
windows-cu113

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant