Skip to content

Qualcomm AI Engine Direct - Decompose integer floor_divide by a Python scalar - #24034

Open
msluszniak wants to merge 1 commit into
pytorch:mainfrom
msluszniak:ms/qnn-floor-divide-scalar
Open

msluszniak wants to merge 1 commit into
pytorch:mainfrom
msluszniak:ms/qnn-floor-divide-scalar

Conversation

@msluszniak

@msluszniak msluszniak commented Oct 21, 2026 •

Copy link
Copy Markdown
Contributor

Summary

DecomposeFloorDivide reads .meta from both operands, so an integer floor_divide by a Python scalar crashes export with 'int' object has no attribute 'meta'. YOLO26's detection head hits this (index // nc). The scalar is now traced as a 0-dim tensor and remapped back.

Test plan

New int_scalar subtest of test_decompose_floor_divide fails on main and passes here; no other result in rework/passes/test.py changes. Standalone, the decomposed x // 3 and x // -4 match eager.

cc @cccclai @winskuo-quic @shewu-quic @haowhsu-quic @DannyYuyang-quic @cbilgin @abhinaykukkadapu @psiddh @nil-is-all

…n scalar

DecomposeFloorDivide read .meta from both operands, so an integer floor_divide by a
Python scalar (YOLO26's `index // nc`) crashed export with "'int' object has no
attribute 'meta'". Trace the scalar as a 0-dim tensor and remap it back, and cast
the result to the node's own dtype.
@msluszniak
msluszniak requested a review from psiddh as a code owner October 21, 2026 05:44
@pytorch-bot

pytorch-bot Bot commented Oct 21, 2026 •

Copy link
Copy Markdown

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24034

Note: Links to docs will display an error until the docs builds have been completed.

❌ 4 New Failures, 2 Unclassified Failures

As of commit 2e6b6d7 with merge base e6cfbf8 (image):

NEW FAILURES - The following jobs have failed:

UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@meta-cla meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Oct 21, 2026
@github-actions

Copy link
Copy Markdown

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

@executorch-triage executorch-triage Bot added the community: contribution PRs coming from community (excluding hardware partners) label Oct 21, 2026
@nil-is-all nil-is-all added partner: qualcomm For backend delegation, kernels, demo, etc. from the 3rd-party partner, Qualcomm module: qnn Issues related to Qualcomm's QNN delegate and code under backends/qualcomm/ labels Oct 21, 2026
borys3kk pushed a commit to borys3kk/swm_executorch that referenced this pull request Oct 22, 2026
…n scalar

DecomposeFloorDivide read .meta from both operands, so an integer floor_divide by a
Python scalar (YOLO26's `index // nc`) crashed export with "'int' object has no
attribute 'meta'". Trace the scalar as a 0-dim tensor and remap it back, and cast
the result to the node's own dtype.

Backport of pytorch/executorch#24034. 1.4.1 has no rework pass
tests, so the subtest is an end-to-end lowering test in test_passes.py here.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. community: contribution PRs coming from community (excluding hardware partners) module: qnn Issues related to Qualcomm's QNN delegate and code under backends/qualcomm/ partner: qualcomm For backend delegation, kernels, demo, etc. from the 3rd-party partner, Qualcomm

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants