-
-
Notifications
You must be signed in to change notification settings - Fork 26.2k
ENH/FIX stopping criterion for coordinate descent gap <= tol
#31906
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
Conversation
* tol <= gap instead of tol < gap
tol <= gap
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.
LGTM. Thanks @lorentzenchr
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.
Thanks for PR, @lorentzenchr!
Just a few minor issues with the rendered docs:

Otherwise, LGTM!
We also have a few other PRs updating _coordinate_descent.py
(e.g., #31882, #31909), so at some point we might need to sort out the merge order.
tol <= gap
gap <= tol
I had missed the parameter docstring of |
This one first, please. |
Technically, this PR has the approval of two core contributors. @OmarManzoor wanna merge? |
doc/whats_new/upcoming_changes/sklearn.linear_model/31906.enhancement.rst
Outdated
Show resolved
Hide resolved
…ncement.rst Co-authored-by: Omar Salman <omar.salman2007@gmail.com>
LGTM. Let's enable auto-merge |
Reference Issues/PRs
Noted already in #20144.
What does this implement/fix? Explain your changes.
The stopping criterion in the main branch reads:
gap < tol
, but a better version isgap <= tol
.Any other comments?
Reduce (or explicitly suppress) many Convergence warnings in test suite, see #4780.