[MRG] Sinkhorn gradient last step #693
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 ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #693 +/- ##
==========================================
+ Coverage 97.05% 97.06% +0.01%
==========================================
Files 98 98
Lines 19877 19955 +78
==========================================
+ Hits 19292 19370 +78
Misses 585 585
|
rflamary
reviewed
Nov 19, 2024
Co-authored-by: Rémi Flamary <[email protected]>
rflamary
reviewed
Nov 20, 2024
| potentials = (log["log_u"], log["log_v"]) | ||
|
|
||
| if grad == "last_step": | ||
| loga = nx.log(a0) |
There was a problem hiding this comment.
maybe a few comments do dicuss the sinkhorn iettaration here
| @@ -412,7 +412,10 @@ def solve( | |||
| potentials = (log["u"], log["v"]) | |||
|
|
|||
| elif reg_type.lower() in ["entropy", "kl"]: | |||
| if grad == "envelope": # if envelope then detach the input | |||
| if grad in [ | |||
| "envelope", | |||
There was a problem hiding this comment.
could you also add "detach" option here and in teh documengation? seems like a nice feature that can be practical
rflamary
changed the title
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in 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Outdated suggestions cannot be applied.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.
Types of changes
I added a
grad=last_stepfeature to theot.solvers.solvefunction to limit the gradient computation to the last iteration of Sinkhorn instead of computing the gradient for every iteration.Motivation and context / Related issue
This change is required in the case where computing the gradient for all the Sinkhorn iterations is too heavy.
How has this been tested (if it applies)
I created a
test_solve_last_steptest to check that the gradient is different when usinggrad=last_stepandgrad=autodiff.PR checklist