-
Notifications
You must be signed in to change notification settings - Fork 4
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
MNT Update dependencies #118
Conversation
…around drop of atomics on CPU to keep a working CI (by pinning CI on an old working version).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If it works, LGTM :)
# Version of other python packages explicitly installed either within the | ||
# build environment or the runtime environment. | ||
|
||
ARG NUMPY_VERSION="<1.25" | ||
ARG NUMBA_VERSION="" | ||
ARG CYTHON_VERSION="<3" | ||
ARG CYTHON_VERSION="" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If Cython is used to build scikit-learn, note that there are a few performance regressions with Cython 3. So we might want to stick to the latest Cython 2 for the time being.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
cython is not installed in the final image, so it's not related to the cython one can use to build sklearn later on in a container. Most recent dpctl has merged work that makes it compatible with cython>=3 so I think we can keep that.
Apparently it doesn't work:
|
It's normal that the test don't pass yet, we must merge first so that the runner build and update the image to our repo, and re-run the pipelines after that. I'll perform a few final tests locally before. |
TY for review |
Update Dockerfile, patch around inline optimization issues, and hack around drop of atomics on CPU to keep a working CI (by pinning CI on an old working version).