Qualcomm AI Engine Direct - Adding QNN backend support for rand core ATen op#18298
Open
qti-horodnic wants to merge 1 commit intopytorch:mainfrom
Open
Qualcomm AI Engine Direct - Adding QNN backend support for rand core ATen op#18298qti-horodnic wants to merge 1 commit intopytorch:mainfrom
qti-horodnic wants to merge 1 commit intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18298
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 7ef4121 with merge base a5423eb ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
Contributor
Author
|
@pytorchbot label "release notes: qualcomm" |
pytorch-bot
bot
added
the
release notes: qualcomm
Contributor
|
@qti-horodnic please rebase this as well. |
| self, | ||
| node: torch.fx.Node, | ||
| nodes_to_wrappers: Dict[torch.fx.Node, PyQnnManager.TensorWrapper], | ||
| ) -> PyQnnManager.PyQnnOpWrapper: |
Contributor
There was a problem hiding this comment.
Nit: Add input type guard to support only int32?
cccclai
reviewed
Mar 19, 2026
Contributor
There was a problem hiding this comment.
Is this number still correct?
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Added support for the nonZero core ATen op using the existing QNN backend implementation of
randomuniformlike. Note only UINT32 is supported, so there is only a quantized test which uses the addition op in order to test the quantized input.Test plan
python backends/qualcomm/tests/test_qnn_delegate.py -k TestQNNQuantizedOperator.test_qnn_backend_rand --model SM8750 --host aisw-vm15-labsd --device 545ee4aa --build_folder build-android