Add dependency workaround for Intel-based Macs#227
Conversation
|
Thank you for opening this PR. Since this is a release branch, the PR must complete the release checklist, below:
|
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## release/v1.8 #227 +/- ##
===============================================
Coverage ? 83.69%
===============================================
Files ? 103
Lines ? 10961
Branches ? 0
===============================================
Hits ? 9174
Misses ? 1787
Partials ? 0 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
ba9fedb to
be447d0
Compare
NOTE: This is a backup option in case the transition from
tensorflowtoPyTorchdoesn't worktensorflowis needed to enable data-driven features in Progpy. However, we need a dependency workaround using an older version oftensorflowat2.16.2since newer versions like2.18.0are not available for Intel-based Macs. Older versions oftensorflowmay have issues with newer versions ofnumpyandnumpoly.The workaround thus includes a know working combination of of
tensorflow==2.16.2,numpy==1.26.4, andnumpoly==1.2.12.More information on
tensorflowsupport, refer to tensorflow/tensorflow#78836