Add a model for probe variation over time with coherent probe modes to the least-squares algorithm and fix a problem with phase ramping at the edges of the reconstruction.
Approach
Please read new documentation in tike.ptycho.probe for information about the variation correction.
To fix the phase ramping problem, I re-implemented the update directions in the least squares method to use weighted sums instead of normalization by division. The one side effect is that the least-squares method seems to converge more slowly now. I will have to see if momentum acceleration can improve convergence rates.
Forced orthogonalization for incoherent probes is now disabled in the least-squares method because it was causing artifacts in my tests.
Pre-Merge Checklists
Submitter
[x] Write a helpfully descriptive pull request title.
[x] Organize changes into logically grouped commits with descriptive commit messages.
[x] Document all new functions.
[x] Write tests for new functions or explain why they are not needed.
Purpose
Add a model for probe variation over time with coherent probe modes to the least-squares algorithm and fix a problem with phase ramping at the edges of the reconstruction.
Approach
Please read new documentation in
tike.ptycho.probefor information about the variation correction.To fix the phase ramping problem, I re-implemented the update directions in the least squares method to use weighted sums instead of normalization by division. The one side effect is that the least-squares method seems to converge more slowly now. I will have to see if momentum acceleration can improve convergence rates.
Forced orthogonalization for incoherent probes is now disabled in the least-squares method because it was causing artifacts in my tests.
Pre-Merge Checklists
Submitter
yapfto format python code.Reviewer