[ONNX] Add eliminate_nop_cast pass #3376
Merged
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
andrey-churkin
commented
Mar 25, 2025
andrey-churkin
commented
Mar 25, 2025
andrey-churkin
commented
Mar 25, 2025
Contributor
Author
|
@kshpv Please review |
andrey-churkin
force-pushed
the
ac/eliminate_nop_cast
branch
from
4650dc4 to
fa04951
Compare
alexsu52
reviewed
Apr 21, 2025
alexsu52
left a comment
alexsu52
left a comment
There was a problem hiding this comment.
Is this PR in draft or can it be reviewed?
andrey-churkin
force-pushed
the
ac/eliminate_nop_cast
branch
from
76053b7 to
ac7b3c2
Compare
andrey-churkin
force-pushed
the
ac/eliminate_nop_cast
branch
from
f059315 to
17f658e
Compare
Contributor
Author
|
NNCF/nightly/e2e_onnx_model_zoo: Build # 953 |
Contributor
Author
I've updated it, so it can now be reviewed. |
Contributor
There was a problem hiding this comment.
Pull Request Overview
This pull request introduces the eliminate_nop_cast optimization pass to remove no-op cast nodes from ONNX graphs before quantization. Key changes include:
- Adding a new pass in nncf/onnx/graph/passes.py to eliminate redundant Cast nodes.
- Integrating the pass into model preprocessing in both the test and quantization flows.
- Creating a test model with a no-op Cast and a corresponding unit test in tests/onnx.
Reviewed Changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| tests/onnx/test_passes.py | Added a unit test to validate the elimination of no-op Cast nodes. |
| tests/onnx/models.py | Introduced a helper to build a MatMul model with a no-op Cast node. |
| nncf/onnx/quantization/quantize_model.py | Integrated preprocess passes into the quantization pipeline. |
| nncf/onnx/graph/passes.py | Implemented the eliminate_nop_cast and apply_preprocess_passes passes. |
Comments suppressed due to low confidence (2)
nncf/onnx/graph/passes.py:58
- Consider verifying that cast_node.output[0] exists in model.graph.value_info before attempting removal, to prevent potential KeyError if the value info is absent.
model.graph.value_info.remove(value_infos[cast_node.output[0]])
tests/onnx/models.py:1911
- Ensure that numpy is imported as 'np' in this module if it is not already imported to avoid a NameError.
W_values = np.random.rand(3, 2).astype(np.float32)
anzr299
reviewed
May 2, 2025
anzr299
approved these changes
May 2, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Changes
Apply the
eliminate_nop_castoptimization pass before building the NNCF graph. This pass aims to eliminate no-op cast nodes in the graph. A no-op cast is a cast that doesn't change the value, such as casting a tensor to its own type.Reason for changes
Statistics cannot be collected after the no-op Cast nodes because such nodes are removed from the ONNX inference graph during the session.
Related tickets
Ref: 164211