CI tests only run on conda
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 17.9k
- Forks
- 7.3k
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 13
Description
🚀 Feature
Run CI tests on pip as well.
Motivation
We have traditionally built torchvision via both conda and pip, but we only perform tests on conda.
This has been generally fine because there was nothing which was specific to pip that could affect runtime and wouldn't be catch by a conda test.
This is not true anymore since the addition of libpng / libjpeg / ffmpeg as dependencies to torchvision, which requires different manipulations for pip.
This lack of coverage is illustrated in https://github.com/pytorch/vision/pull/2775, where only the conda binaries work, and not the pip packages.
cc @andfoy @seemethere
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing the CI configuration and packaging/build_wheel.sh, especially the pip-specific dependency handling described in the issue. Compare the existing conda test flow with the pip package path and use PR 2775 as a reproduction reference. Done means CI installs the pip package and runs the relevant tests successfully.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- build-system, ci-cd, testing-qa
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100