Fix Issue #2583: Dynamic padding in torch.nn.functional.pad #2588
Conversation
This file contains hidden or 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
Modified _array_construct to handle dynamic padding values: Creates proper Var objects using mb.concat instead of Python lists + Fixes AttributeError when converting models with x.size(-1) padding
|
Hey @karimnosseir, can you review my PR please? |
|
@noobsiecoder Thank you for the contribution! Let me kick off a CI run for this PR. |
|
@noobsiecoder There are some failures in the CI link above. Could you take a look to fix it? Thanks! |
|
@junpeiz, yes I'm on it. |
Co-authored-by: yifan_shen3 <[email protected]>
…2585) (apple#2592) * Add RMSNorm operator support for PyTorch to CoreML conversion (apple#2585) * formatted code * handles FP16 overflow for RMSNorm operation
|
Hi @junpeiz, I found the issue and fixed it. Could you please try running CI now? |
|
Sure! Started another CI: https://gitlab.com/coremltools1/coremltools/-/pipelines/2033126061 |
junpeiz
approved these changes
Sep 11, 2025
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Outdated suggestions cannot be applied.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.
Description
Fixes #2583
This PR fixes the conversion of
torch.nn.functional.padoperations with dynamic padding values (e.g.,x.size(-1)) to Core ML format.Problem
When padding values contain runtime-determined sizes, the
_array_constructfunction inops.pywas returning a Python list instead of a proper MILVarobject, causingAttributeError: 'list' object has no attribute 'val'.Solution
Modified
_array_constructto create a properVarobject usingmb.concatfor dynamic list construction, maintaining the expectedVarinterface throughout the conversion pipeline.Testing
test_ops.pycovering dynamic, constant, and mixed padding casesChecklist for Reviewer(s)