Add RMSNorm operator support for PyTorch to CoreML conversion (#2585) #2592
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
|
Great! Will merge after getting a green CI: https://gitlab.com/coremltools1/coremltools/-/pipelines/2030304960 |
junpeiz
approved these changes
Sep 10, 2025
|
Hey @junpeiz, can I merge my changes to the main branch? |
No worries! I already merged it. Thank you for your contributions! |
noobsiecoder
added a commit
to noobsiecoder/coremltools
that referenced
this pull request
Sep 10, 2025
…2585) (apple#2592) * Add RMSNorm operator support for PyTorch to CoreML conversion (apple#2585) * formatted code * handles FP16 overflow for RMSNorm operation
|
Relevant to your discussion about OpenELM: huggingface/swift-transformers#95 |
junpeiz
pushed a commit
that referenced
this pull request
Sep 11, 2025
* Fix Issue #2583: Dynamic padding in torch.nn.functional.pad 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 * limit torch to older than 2.8 for now (#2591) Co-authored-by: yifan_shen3 <[email protected]> * Add RMSNorm operator support for PyTorch to CoreML conversion (#2585) (#2592) * Add RMSNorm operator support for PyTorch to CoreML conversion (#2585) * formatted code * handles FP16 overflow for RMSNorm operation * handle dynamic padding w/o breaking legacy code --------- Co-authored-by: Yifan Shen <[email protected]> Co-authored-by: yifan_shen3 <[email protected]>
|
thanks for the tag @FL33TW00D. i'd switch to using anemll rms norm implementation, it gives me the impression that it relies on ane functionalities not publicly accesible. |
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 #2585
Added PyTorch nn.RMSNorm to CoreML conversion by decomposing into primitive MIL operations.
Implementation
Decompose RMSNorm formula into
mbcomposable functions:References
Checklist for Reviewer(s)
coremltools/converters/mil/test/test_rms_norm.py