Repository navigation
Allow CoordinateDescent 0.6, version 0.2.3 - #6
Merged
Merged
Conversation
CoordinateDescent 0.6 uses momentum by default with SparseIterates and its built-in losses. In DiffPrecTest that covers reducedDiffEstimation and invHessianReduced, which use CDQuadraticLoss, and the nodewise lasso fits; their solutions agree with 0.5 up to the tolerance. The other estimators use DiffPrecTest's own losses or SymmetricSparseIterates and run as before. Co-Authored-By: Claude Opus 5.5 <[email protected]>
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #6 +/- ##
=======================================
Coverage 98.65% 98.65%
=======================================
Files 12 12
Lines 1409 1409
=======================================
Hits 1390 1390
Misses 19 19 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merge after mlakolar/CovarianceSelection.jl#15. CI takes CovSel
masterthrough[sources], and throughPkg.add(url=...)on Julia 1.10. CovSel master allows only CoordinateDescent"0.4, 0.5"until that PR merges, so until then this PR's CI resolves CoordinateDescent 0.5.6. Re-run CI after CovSel#15 merges to test 0.6.CoordinateDescent 0.6.0 (mlakolar/HighDimensionalModels.jl#22, registered in General #173025) uses Nesterov momentum by default with a
SparseIterate, a random order and one of its built-in losses. In DiffPrecTest that covers:reducedDiffEstimationandinvHessianReduced, which useCDQuadraticLoss;lassofits.Their solutions agree with 0.5 up to the tolerance, not bit for bit. The other estimators run as before:
CDInverseKroneckerLoss,CDInverseReducedSymKroneckerLossandCDDirectDifferenceLossare DiffPrecTest's own losses;diffEstimationand the support estimates useSymmetricSparseIterates.This PR widens the compat to
"0.4, 0.5, 0.6"and bumps the version to 0.2.3.Testing: the test suite (20186 tests) passed against the CoordinateDescent 0.6 branch, using a CovSel that allows 0.6.
🤖 Generated with Claude Code