Skip to content

Conversation

NicolasHug
Copy link
Contributor

I think we shouldn't need torchvision since #612

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Meta Open Source bot. label Apr 22, 2025
- name: Install PyTorch
run: |
${CONDA_RUN} python -m pip install --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cu${{ env.cuda_version_without_periods }}
${CONDA_RUN} python -m pip install --pre torch --index-url https://download.pytorch.org/whl/nightly/cu${{ env.cuda_version_without_periods }}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like we still need torchvision here because of the GPU benchmarks.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

aaaaah I forgot we run those. We probably... shouldn't, since we never check 🤷

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think there's some value in doing it, as it makes sure that the GPU benchmarks don't become stale. That is, it's not about the actual performance result from the benchmark, it's more a check that it's still usable. There may also be value in doing that on the CPU side as well.

@scotts
Copy link
Contributor

scotts commented Apr 22, 2025

Current CI failures look like the pre-existing problem, so this should improve things.

@NicolasHug NicolasHug merged commit 879cae2 into meta-pytorch:main Apr 23, 2025
48 of 50 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Meta Open Source bot.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants