-
Notifications
You must be signed in to change notification settings - Fork 416
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
[RFC] Use uv
package manager in GHA
#2713
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2713 +/- ##
=======================================
Coverage 99.98% 99.98%
=======================================
Files 202 202
Lines 18602 18602
=======================================
Hits 18600 18600
Misses 2 2 ☔ View full report in Codecov by Sentry. |
sure that sounds like a good change, let's update this across the board. |
@saitcakmak has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
@saitcakmak merged this pull request in 4a595f2. |
Summary: Replicates pytorch/botorch#2713 for Ax. `uv` is much faster than `pip`, which leads to a couple minutes of speed up in each workflow, which can add up to meaningful compute savings over many CI runs we do every day. Pull Request resolved: #3290 Test Plan: Successful CI run Compared to a previous CI run Before - 155 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13040634315/job/36381474759 After - 48 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13061685440/job/36445887187 Reviewed By: Balandat Differential Revision: D68919180 Pulled By: saitcakmak
Summary: Replicates pytorch/botorch#2713 for Ax. `uv` is much faster than `pip`, which leads to a couple minutes of speed up in each workflow, which can add up to meaningful compute savings over many CI runs we do every day. Pull Request resolved: #3290 Test Plan: Successful CI run Compared to a previous CI run Before - 155 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13040634315/job/36381474759 After - 48 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13061685440/job/36445887187 Reviewed By: Balandat Differential Revision: D68919180 Pulled By: saitcakmak
Summary: Replicates pytorch/botorch#2713 for Ax. `uv` is much faster than `pip`, which leads to a couple minutes of speed up in each workflow, which can add up to meaningful compute savings over many CI runs we do every day. Pull Request resolved: #3290 Test Plan: Successful CI run Compared to a previous CI run Before - 155 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13040634315/job/36381474759 After - 48 secs of setup time (until test & coverage step): https://github.com/facebook/Ax/actions/runs/13061685440/job/36445887187 Reviewed By: Balandat Differential Revision: D68919180 Pulled By: saitcakmak fbshipit-source-id: 3953ce5ff1aaae78587e56ab4046f70fee86cdf7
I came across
uv
while debugging the scipy slowdowns and noticed how much faster it is than vanillapip
. Usinguv
saves a couple minutes in installing dependencies, which can add up to meaningful compute savings over many CI runs we do every day. Should we update all workflows to useuv
?This PR updates
reusable_tutorials.yml
as an example.Before, with
pip
: https://github.com/pytorch/botorch/actions/runs/13058152281/job/36434279975After, with
uv pip
: https://github.com/pytorch/botorch/actions/runs/13058212963/job/36434483617