fix(MemoryFormatOpsPass): preserve input dim_order for clone/to_copy with no memory_format kwarg by nefainl · Pull Request #17611 · pytorch/executorch

nefainl · 2026-02-21T11:32:37Z

Summary

This PR fixes MemoryFormatOpsPass to correctly handle torch.preserve_format semantics for clone() and _to_copy.default operations.

Root cause: When clone() or _to_copy is called without an explicit memory_format kwarg, the pass was defaulting to torch.contiguous_format, causing the output dim_order to be [0,1,2,3] (contiguous) even when the input was channels-last [0,2,3,1]. This caused runtime assertion failures:

Code=18 InvalidArgument: tensors_have_same_dim_order(self, out)

Fix: Change the default from torch.contiguous_format to torch.preserve_format, and derive dim_order from the input tensor's dim_order() when preserve_format is used.

This is a minimal, focused fix following the guidance from @GregoryComer in the discussion on PR #17463.

Changes

exir/passes/memory_format_ops_pass.py (+29/-5 lines):
- Default memory_format to torch.preserve_format instead of torch.contiguous_format
- When preserve_format, derive dim_order from input_tensor.dim_order()
- Fallback to contiguous if no input tensor available (e.g., empty())
exir/tests/test_passes.py (+130 lines):
- test_clone_no_kwarg_preserves_channels_last_dim_order: Core repro case for Dim Order Validation Inconsistency for Edge / Ambiguous Cases #16032
- test_clone_contiguous_format_kwarg_stays_contiguous: Regression guard
- test_to_copy_no_kwarg_preserves_channels_last_dim_order: Verifies _to_copy.default path

Standalone Reproduction

import torch
from torch.export import export
from executorch.exir import to_edge, EdgeCompileConfig

class ConvClone(torch.nn.Module):
    def __init__(self):
        super().__init__()
        self.conv = torch.nn.Conv2d(3, 16, kernel_size=3, padding=1)

    def forward(self, x):
        return self.conv(x).clone()

model = ConvClone().to(memory_format=torch.channels_last)
x = torch.randn(1, 3, 8, 8).to(memory_format=torch.channels_last)

exported = export(model, (x,))
edge = to_edge(exported, compile_config=EdgeCompileConfig(_skip_dim_order=False))

# Before fix: clone node has dim_order=(0,1,2,3) - BUG
# After fix: clone node has dim_order=(0,2,3,1) - CORRECT
for node in edge.exported_program().graph_module.graph.nodes:
    if "_clone_dim_order" in str(node.target):
        print(f"clone dim_order: {tuple(node.meta['val'].dim_order())}")

Test Plan

All 3 new tests pass
Verified fix with standalone reproduction script
No changes to existing tests required

…with no memory_format kwarg Issue pytorch#16032: clone() and _to_copy operations with no explicit memory_format kwarg were defaulting to contiguous dim_order, causing runtime assertion failures when cloning channels-last tensors. Changes: - Default memory_format to torch.preserve_format instead of torch.contiguous_format - When preserve_format, derive dim_order from input tensor's dim_order() - Simplify type annotation: dim_order is always assigned, no Optional needed Tests: - test_clone_no_kwarg_preserves_channels_last_dim_order: core repro case - test_clone_contiguous_format_kwarg_stays_contiguous: regression guard - test_to_copy_no_kwarg_preserves_channels_last_dim_order: _to_copy path

pytorch-bot · 2026-02-21T11:32:41Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17611

📄 Preview Python docs built from this PR

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

❌ 2 New Failures

As of commit 4f91bc4 with merge base 9f2f005 ():

NEW FAILURES - The following jobs have failed:

pull / test-samsung-models-linux / linux-job (gh)
RuntimeError: Command docker exec -t 5d68fafcf43d5fe01e6be445b053b20a039dc5fa55eac8d6b31b31d0d2b92001 /exec failed with exit code 1
pull / test-samsung-quantmodels-linux / linux-job (gh)
RuntimeError: Command docker exec -t 7529190d67318384672aec373aaa41c86dd21dddaae686c9dcb9d3147f0f82a3 /exec failed with exit code 1

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

nefainl · 2026-02-21T11:38:26Z

@pytorchbot label "release notes: exir"

nefainl · 2026-02-21T11:50:34Z

Request to the reviewers: Could you please add @GregoryComer as well? I have received very helpful advice from him in the other PR and it makes sense to have him also here so he can edit if needed. Note that some of the other changes will still be handled in another PR (where it also makes sense to have him review).

…rmat

GregoryComer

@nefainl Thanks for the update. This PR looks good, we can merge with a few small changes.

In addition to the few small nits, can you update the tests to also run the model and verify that the output tensor has the correct memory format / dim order?

nefainl · 2026-02-26T21:33:35Z

@nefainl Thanks for the update. This PR looks good, we can merge with a few small changes.

In addition to the few small nits, can you update the tests to also run the model and verify that the output tensor has the correct memory format / dim order?

@GregoryComer Thank you for looking into it, I will implement your suggestions and look into the lint runner output tomorrow and make the required fixes!

nefainl · 2026-02-26T21:41:50Z

Correction, I will look in detail at all the failing checks to see how to resolve them tomorrow.

GregoryComer · 2026-02-26T21:44:04Z

Correction, I will look in detail at all the failing checks to see how to resolve them tomorrow.

Thanks. You can ignore the samsung failures. The other failures (except lint) look like flakes. I'll re-try the jobs.

nefainl · 2026-02-27T20:02:53Z

Quick update on this PR:

Updated MemoryFormatOpsPass so that:
- memory_format defaults to torch.preserve_format instead of torch.contiguous_format.
- For preserve-format (None or torch.preserve_format), dim_order is now derived from the input tensor’s dim_order() when available.
- When there is no single input tensor (e.g. immutable list inputs for ops like torch.stack), we conservatively fall back to contiguous dim_order = list(range(ndim)).
Kept the existing behavior for explicit memory_format values (contiguous, channels_last, etc.), which still go through get_dim_order(mem_format, ndim).
Addressed the style nits by removing explicit GitHub issue references from comments/docstrings and keeping them in the commit/PR description instead.

Locally, the new TestMemoryFormatOpsPassPreserveFormat tests pass and confirm:

clone() with no memory_format kwarg preserves channels-last dim_order in the exported graph.
clone(memory_format=torch.contiguous_format) produces a contiguous layout.
to(dtype=...) without memory_format preserves channels-last dim_order.

Let me know if you’d like any additional end-to-end checks (e.g. running the exported graph and inspecting runtime dim_order) beyond what’s already covered in the tests.

…erve-format

Clarify preserve_format behavior and extend MemoryFormatOpsPass tests to run the models and assert that output tensors have the expected memory format / dim order.

Replace the generator-based dim_order construction with a list comprehension to satisfy FLAKE8 C400 and add the missing blank line before the new test class to align with PEP 8 spacing.

nefainl · 2026-02-27T20:54:32Z

Correction, I will look in detail at all the failing checks to see how to resolve them tomorrow.

Thanks. You can ignore the samsung failures. The other failures (except lint) look like flakes. I'll re-try the jobs.

Thanks a lot for the input and the helpful review comments / nits etc. I think I have covered all now. If you see any additional items please let me know!

nefainl · 2026-02-27T20:59:02Z

Each of the three TestMemoryFormatOpsPassPreserveFormat tests now:

Runs the model eagerly with the appropriate input (channels-last or contiguous as needed).
Asserts the output tensor’s memory format via is_contiguous() / is_contiguous(memory_format=torch.channels_last) so we verify the actual tensor layout, not only the exported graph.
Then exports and checks that the graph’s dim_order metadata matches (e.g. (0, 2, 3, 1) for channels-last, (0, 1, 2, 3) for contiguous).

So we’re already checking both runtime output layout and export metadata. If you’d like additional end-to-end checks (e.g. running the exported ExecuTorch program and inspecting runtime dim_order), we can add those.

…rmat

GregoryComer · 2026-03-05T18:14:49Z

Thanks - looks good. I'm going to run a few more tests and I'll merge if everything looks good.

meta-codesync · 2026-03-05T18:18:26Z

@GregoryComer has imported this pull request. If you are a Meta employee, you can view this in D95409924.

nefainl · 2026-03-05T21:46:48Z

Thanks - looks good. I'm going to run a few more tests and I'll merge if everything looks good.

Great, thanks for your help earlier in the solution direction! I will look into the other PR in more detail later.

@GregoryComer

…with no memory_format kwarg (pytorch#17611) ## Summary Fixes pytorch#16032 This PR fixes `MemoryFormatOpsPass` to correctly handle `torch.preserve_format` semantics for `clone()` and `_to_copy.default` operations. **Root cause:** When `clone()` or `_to_copy` is called without an explicit `memory_format` kwarg, the pass was defaulting to `torch.contiguous_format`, causing the output `dim_order` to be `[0,1,2,3]` (contiguous) even when the input was channels-last `[0,2,3,1]`. This caused runtime assertion failures: ``` Code=18 InvalidArgument: tensors_have_same_dim_order(self, out) ``` **Fix:** Change the default from `torch.contiguous_format` to `torch.preserve_format`, and derive `dim_order` from the input tensor's `dim_order()` when preserve_format is used. This is a minimal, focused fix following the guidance from @GregoryComer in the discussion on PR pytorch#17463. ## Changes - **`exir/passes/memory_format_ops_pass.py`** (+29/-5 lines): - Default `memory_format` to `torch.preserve_format` instead of `torch.contiguous_format` - When preserve_format, derive `dim_order` from `input_tensor.dim_order()` - Fallback to contiguous if no input tensor available (e.g., `empty()`) - **`exir/tests/test_passes.py`** (+130 lines): - `test_clone_no_kwarg_preserves_channels_last_dim_order`: Core repro case for pytorch#16032 - `test_clone_contiguous_format_kwarg_stays_contiguous`: Regression guard - `test_to_copy_no_kwarg_preserves_channels_last_dim_order`: Verifies `_to_copy.default` path ## Standalone Reproduction ```python import torch from torch.export import export from executorch.exir import to_edge, EdgeCompileConfig class ConvClone(torch.nn.Module): def __init__(self): super().__init__() self.conv = torch.nn.Conv2d(3, 16, kernel_size=3, padding=1) def forward(self, x): return self.conv(x).clone() model = ConvClone().to(memory_format=torch.channels_last) x = torch.randn(1, 3, 8, 8).to(memory_format=torch.channels_last) exported = export(model, (x,)) edge = to_edge(exported, compile_config=EdgeCompileConfig(_skip_dim_order=False)) # Before fix: clone node has dim_order=(0,1,2,3) - BUG # After fix: clone node has dim_order=(0,2,3,1) - CORRECT for node in edge.exported_program().graph_module.graph.nodes: if "_clone_dim_order" in str(node.target): print(f"clone dim_order: {tuple(node.meta['val'].dim_order())}") ``` ## Test Plan - [x] All 3 new tests pass - [x] Verified fix with standalone reproduction script - [x] No changes to existing tests required ## Related - Fixes pytorch#16032 - Supersedes pytorch#17463 (this is the minimal fix extracted from that PR per reviewer feedback) --------- Co-authored-by: NefAI <info@nefai.nl>

nefainl requested review from JacobSzwejbka and larryliu0820 as code owners February 21, 2026 11:32

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 Feb 21, 2026

pytorch-bot Bot added the release notes: exir Changes to any dialects and passes on these dialects, such as memory planning label Feb 21, 2026

nefainl mentioned this pull request Feb 21, 2026

Fix #16032: propagate channels_last dim_order to out-variant TensorSp… #17463

Draft

nefainl mentioned this pull request Feb 21, 2026

Fix tensors_have_same_dim_order for degenerate shapes (semantic equivalence) #17612

Open

4 tasks

Merge branch 'main' into fix/16032-memory-format-ops-pass-preserve-fo…

33eb3e1

…rmat

mergennachin requested review from GregoryComer, psiddh and rascani February 23, 2026 17:33

nefainl added 3 commits February 24, 2026 20:35

Merge branch 'main' into fix/16032-memory-format-ops-pass-preserve-fo…

377a635

…rmat

Merge branch 'main' into fix/16032-memory-format-ops-pass-preserve-fo…

8896c78

…rmat

Merge branch 'main' into fix/16032-memory-format-ops-pass-preserve-fo…

faba24a

…rmat

GregoryComer requested changes Feb 26, 2026

View reviewed changes

Comment thread exir/passes/memory_format_ops_pass.py Outdated

Comment thread exir/tests/test_passes.py Outdated

Comment thread exir/tests/test_passes.py Outdated

NefAI added 4 commits February 27, 2026 21:12

chore: merge upstream main into fix/16032-memory-format-ops-pass-pres…

daebebf

…erve-format

fix: update MemoryFormatOpsPass comments and tests

45ab83a

Clarify preserve_format behavior and extend MemoryFormatOpsPass tests to run the models and assert that output tensors have the expected memory format / dim order.

fix: address lint and style for MemoryFormatOpsPass

c1814d6

Replace the generator-based dim_order construction with a list comprehension to satisfy FLAKE8 C400 and add the missing blank line before the new test class to align with PEP 8 spacing.

docs: clarify fallback comment and PEP 8 spacing before TestCSEPass

0ebd613

nefainl requested a review from GregoryComer February 27, 2026 20:59

Merge branch 'main' into fix/16032-memory-format-ops-pass-preserve-fo…

4f91bc4

…rmat

GregoryComer approved these changes Mar 5, 2026

View reviewed changes

GregoryComer merged commit 4c20ef1 into pytorch:main Mar 5, 2026
154 of 157 checks passed

Conversation

nefainl commented Feb 21, 2026

Summary

Changes

Standalone Reproduction

Test Plan

Related

Uh oh!

pytorch-bot Bot commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17611

❌ 2 New Failures

Uh oh!

nefainl commented Feb 21, 2026

Uh oh!

nefainl commented Feb 21, 2026

Uh oh!

GregoryComer left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

nefainl commented Feb 26, 2026

Uh oh!

nefainl commented Feb 26, 2026

Uh oh!

GregoryComer commented Feb 26, 2026

Uh oh!

nefainl commented Feb 27, 2026

Uh oh!

nefainl commented Feb 27, 2026

Uh oh!

nefainl commented Feb 27, 2026

Uh oh!

GregoryComer commented Mar 5, 2026

Uh oh!

meta-codesync Bot commented Mar 5, 2026

Uh oh!

nefainl commented Mar 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

pytorch-bot Bot commented Feb 21, 2026 •

edited

Loading