Skip to content
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

FIX: Use lsqr solver for spline fit, rerun on extreme values #366

Merged
merged 2 commits into from
Jun 8, 2023

Conversation

effigies
Copy link
Member

@effigies effigies commented Jun 8, 2023

Closes #365.

@effigies effigies added this to the 2.5.1 milestone Jun 8, 2023
effigies added a commit to effigies/fmriprep that referenced this pull request Jun 8, 2023
Seems like it can be slower in some cases, but doesn't have the >30min
worst case performance.
@effigies effigies changed the title FIX: Rerun bspline fit 3 times, check for extreme values FIX: Use lsqr solver for spline fit, rerun on extreme values Jun 8, 2023
@effigies
Copy link
Member Author

effigies commented Jun 8, 2023

@mgxd @oesteban This node is run for both ds054 and ds210 in nipreps/fmriprep#3031. This drops the ds210 runtime of the node from >2000s (when it fails to converge) to 13s. ds054 runs in 4s. I'm keeping in the rerun in case lsqr also has some edge cases, but I think it's a straight win and fixes fMRIPrep CI.

@oesteban
Copy link
Member

oesteban commented Jun 8, 2023 via email

Copy link
Contributor

@mgxd mgxd left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

awesome!

@effigies effigies merged commit 34ae6f7 into nipreps:master Jun 8, 2023
@effigies effigies deleted the fix/rerun-bspline branch June 8, 2023 19:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Ridge regression in BSplineApprox can randomly produce garbage
3 participants