Gauss Newton & LM Robustness Fixes #258
Merged
Conversation
This file contains 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
Codecov Report
@@ Coverage Diff @@
## master #258 +/- ##
==========================================
- Coverage 94.48% 92.47% -2.02%
==========================================
Files 18 18
Lines 1614 1687 +73
==========================================
+ Hits 1525 1560 +35
- Misses 89 127 +38
... and 2 files with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
b25c8bc to
80e6459
Compare
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
6e5b422 to
fabd33c
Compare
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
2 times, most recently
from
9d81e46 to
b4ea93f
Compare
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
93e2815 to
2fdc122
Compare
avik-pal
commented
Oct 23, 2023
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
6 times, most recently
from
997fb49 to
417701c
Compare
src/utils.jl
Outdated
| NormalBunchKaufmanFactorization) | ||
| @eval needs_square_A(::$(alg)) = false | ||
| end | ||
| for kralg in (LinearSolve.Krylov.lsmr!, LinearSolve.Krylov.craigmr!) |
There was a problem hiding this comment.
GMRES requires square matrix
There was a problem hiding this comment.
That's interesting 😅 . Algorithmically that's not required. It is an algorithm that provably gives the L2 solution on non-square matrices, just like QR.
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
22b7f25 to
0c1e739
Compare
axla-io
reviewed
Oct 23, 2023
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
0c1e739 to
4f632dd
Compare
axla-io
reviewed
Oct 23, 2023
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
2 times, most recently
from
be854ec to
6e0911c
Compare
axla-io
reviewed
Oct 23, 2023
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
3 times, most recently
from
4e78736 to
5e7610a
Compare
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
02cf007 to
0f261a4
Compare
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
4 times, most recently
from
72fd691 to
4f3c047
Compare
|
This needs a rebase? |
avik-pal
force-pushed
the
ap/fix_gauss_newton
branch
from
4f3c047 to
2700fce
Compare
|
I am really unsure what to do with the Broyden/Klement failures. They fail on 1.9 but magically pass on 1.10 |
|
@ChrisRackauckas I am disabling the Broyden and Klement 23 test problem tests. I will open and issue and we can deal with it later. No reason to hold off this PR which solves a completely tangential issue |
# for free
to join this conversation on GitHub.
Already have an account?
# 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 #257
For
GaussNewtondefaults toQRFactorization(ColumnNorm()), to ensure we can directly solve forFor
LMwe solve