fix(ops): Fix inconsistent padding calculation in PyTorch backend ops #20774
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
Was able to still reproduce the error, the PyTorch backend had inconsistent behavior between static shape inference and dynamic execution for pooling operations, particularly with 'same' padding and non-unit strides, figured that the root cause was by incorrect padding calculation logic that didn't properly handle asymmetric padding cases. Key changes: - Rewrote _compute_padding_length() to handle stride-based padding - Fixed padding calculation to properly support asymmetric padding cases - Standardize channels_first/channels_last conversion in pooling ops - Cleaned up padding application in _apply_same_padding() - Added proper handling of data_format throughout pooling pipeline This fixes the issue where MaxPooling2D with 'same' padding would produce different shapes between compute_output_shape() and actual execution (e.g. (1,5,2,2) vs (1,5,2,1)). Rebased on top of Sachin's September 2024 PR to incorporate latest keras:master changes.
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #20774 +/- ##
=======================================
Coverage 81.99% 81.99%
=======================================
Files 555 555
Lines 51805 51808 +3
Branches 8012 8013 +1
=======================================
+ Hits 42475 42482 +7
+ Misses 7378 7376 -2
+ Partials 1952 1950 -2
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
fchollet
approved these changes
Jan 17, 2025
google-ml-butler
bot
added
kokoro:force-run
ready to pull
google-ml-butler
bot
removed
awaiting review
ready to pull
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.
Was able to still reproduce the error, the PyTorch backend had inconsistent behavior between static shape inference and dynamic execution for pooling operations, particularly with 'same' padding and non-unit strides, figured that the root cause was by incorrect padding calculation logic that didn't properly handle asymmetric padding cases.
Key changes:
_compute_padding_length()to handle stride-based paddingchannels_first/channels_lastconversion in pooling ops_apply_same_padding()data_formatthroughout pooling pipelineThis fixes the issue where MaxPooling2D with 'same' padding would produce different shapes between compute_output_shape() and actual execution (e.g. (1,5,2,2) vs (1,5,2,1)) #20235.
Rebased on top of @sachinprasadhs's September 2024 PR to incorporate latest
keras:masterchanges #20270.