implement gradient masking with cropping - new PR #101
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
|
Thanks @JonasHell. I will have a look soon. |
There was a problem hiding this comment.
Thanks for working on this! The changes to loss_wrapper.py all look good to me.
It would be good to extend the test cases a bit to cover all the changes:
- In
test_masking: test for bothcropandmultiplymasking methods that the masked gradients are zero. (Can just put this in a loop whereApplyAndRemoveMaskis called with the corresponding argument.) The test could also be renamed totest_ApplyAndRemove...to make the naming more consistent. - Add a test for
MaskIgnoreLabelalong the lines oftest_masking.
With the extended tests this is good to go from my side.
|
Done :) |
constantinpape
approved these changes
Dec 13, 2022
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Outdated suggestions cannot be applied.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.
I cleaned my git repo. Unfortunately, renaming of branches leads to the closing of connected PRs. So I opened a new one. To not loose the old discussion, I link the old PR here.