Skip to content

Conversation

jangorecki
Copy link
Member

closes #7349

  • use more numerically stable algo for prod fast
  • adaptive prod fast redirects to exact
    while it feels ok for adaptive=T/F to have an impact on accuracy, the partial=T/F should not, and because of how partial is implemented, by redirect to adaptive, and because adaptive prod fast was suffering numerical stability due to cumulative prod (new tests added here also detects that), I decided to remove adaptive prod fast and use exact instead. Tips on possible implementation have been provided in the code.

Copy link

codecov bot commented Sep 28, 2025

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.10%. Comparing base (cf898ab) to head (877dbee).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files
@@            Coverage Diff             @@
##           master    #7351      +/-   ##
==========================================
- Coverage   99.10%   99.10%   -0.01%     
==========================================
  Files          84       84              
  Lines       16128    16114      -14     
==========================================
- Hits        15983    15969      -14     
  Misses        145      145              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@jangorecki jangorecki marked this pull request as draft October 3, 2025 20:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

rolling product and numerical stability
1 participant