Cortex-M: Add MLPerf Tiny model tests#18546
Merged
rascani merged 2 commits intopytorch:mainfrom Apr 2, 2026
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18546
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 3 Unrelated FailuresAs of commit eea1921 with merge base c39a0ff ( NEW FAILURE - The following job has failed:
BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
This PR needs a
|
zingo
added
the
module: microcontrollers
rascani
force-pushed
the
cortex-m-mlperf-tiny-tests
branch
from
eb30d0a to
0468885
Compare
rascani
force-pushed
the
cortex-m-mlperf-tiny-tests
branch
from
0468885 to
95cfdb9
Compare
psiddh
approved these changes
Mar 31, 2026
rascani
requested review from
GregoryComer,
JacobSzwejbka,
SS-JIA,
abhinaykukkadapu,
cccclai,
digantdesai,
kirklandsign,
larryliu0820,
lucylq,
mergennachin,
robert-kalmar and
shoumikhin
as code owners
Add Cortex-M dialect and implementation tests for DeepAutoEncoder, DS-CNN, MobileNet V1 0.25, and ResNet-8 from the MLPerf Tiny benchmark suite. All 4D model inputs use explicit channels_last memory format as required by CMSIS-NN. Add atol parameter to CortexMTester.test_dialect() and test_implementation() for models that need relaxed tolerance. Fix McuTestCase.example_inputs type annotation to accept both tuples and callables, matching actual usage. Co-authored-by: Claude <noreply@anthropic.com>
The dialect test was using atol=0.1 while the implementation test used the default atol=1e-03, which is backwards — the CPU graph comparison should not need a looser tolerance than hardware simulation. Remove the override; the test passes at the stricter default. Co-authored-by: Claude <noreply@anthropic.com>
rascani
force-pushed
the
cortex-m-mlperf-tiny-tests
branch
from
0399baf to
eea1921
Compare
rascani
removed request for
GregoryComer,
JacobSzwejbka,
abhinaykukkadapu,
cccclai,
kirklandsign,
larryliu0820,
lucylq and
robert-kalmar
jpiat
pushed a commit
to jpiat/executorch
that referenced
this pull request
Apr 14, 2026
### Summary Add Cortex-M dialect and implementation tests for DeepAutoEncoder, DS-CNN, MobileNet V1 0.25, and ResNet-8 from the MLPerf Tiny benchmark suite. All 4D model inputs use explicit channels_last memory format as required by CMSIS-NN. Add atol parameter to CortexMTester.test_dialect() and test_implementation() for models that need relaxed tolerance. Fix McuTestCase.example_inputs type annotation to accept both tuples and callables, matching actual usage. Fixes pytorch#16038 ### Test plan ``` pytest backends/cortex_m/test/models/ ``` cc @psiddh @AdrianLundell @digantdesai --------- Co-authored-by: Claude <noreply@anthropic.com>
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
3 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
Add Cortex-M dialect and implementation tests for DeepAutoEncoder, DS-CNN, MobileNet V1 0.25, and ResNet-8 from the MLPerf Tiny benchmark suite. All 4D model inputs use explicit channels_last memory format as required by CMSIS-NN.
Add atol parameter to CortexMTester.test_dialect() and test_implementation() for models that need relaxed tolerance.
Fix McuTestCase.example_inputs type annotation to accept both tuples and callables, matching actual usage.
Fixes #16038
Test plan
cc @psiddh @AdrianLundell @digantdesai