Add U55 and U85 16A8W linear tests #14368
Merged
+71
−0
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
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/) [ghstack-poisoned]
Ninja91
added a commit
that referenced
this pull request
Sep 17, 2025
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/) ghstack-source-id: 310226077 Pull Request resolved: #14368
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/14368
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New Failures, 1 PendingAs of commit 2cb2a6c with merge base 07d1092 ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
per
reviewed
Sep 17, 2025
|
Added |
digantdesai
reviewed
Sep 17, 2025
Yep, unrelated @digantdesai |
digantdesai
approved these changes
Sep 18, 2025
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/) [ghstack-poisoned]
Ninja91
added a commit
that referenced
this pull request
Sep 19, 2025
Pull Request resolved: #14368 Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 ghstack-source-id: 310974295 @exported-using-ghexport Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/)
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/) [ghstack-poisoned]
Ninja91
added a commit
that referenced
this pull request
Sep 19, 2025
Pull Request resolved: #14368 Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 ghstack-source-id: 310975048 @exported-using-ghexport Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/)
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/) [ghstack-poisoned]
Ninja91
added a commit
that referenced
this pull request
Sep 20, 2025
Pull Request resolved: #14368 Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations. This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error. The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms: - test_linear_16a8w_u55_INT16: Tests on U55 with rank1 data - test_linear_16a8w_u85_INT16: Tests on U85 with both rank1 and rank4 data Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947. Addresses: #13729 ghstack-source-id: 310981820 @exported-using-ghexport Differential Revision: [D82618039](https://our.internmc.facebook.com/intern/diff/D82618039/)
facebook-github-bot
merged commit 9e46d9c
into
gh/Ninja91/19/base
278 of 282 checks passed
facebook-github-bot
temporarily deployed
to
cherry-pick-bot
GitHub Actions
Inactive
Ninja91
added a commit
that referenced
this pull request
Sep 21, 2025
This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: #14368 by @Ninja91 ^ Please use this as the source of truth for the PR details, comments, and reviews ghstack PR base: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/head Merge bot PR base: https://github.com/pytorch/executorch/tree/main Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/orig @diff-train-skip-merge Co-authored-by: Nitin Jain <[email protected]>
StrycekSimon
pushed a commit
to nxp-upstream/executorch
that referenced
this pull request
Sep 23, 2025
This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: pytorch#14368 by @Ninja91 ^ Please use this as the source of truth for the PR details, comments, and reviews ghstack PR base: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/head Merge bot PR base: https://github.com/pytorch/executorch/tree/main Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/19/orig @diff-train-skip-merge Co-authored-by: Nitin Jain <[email protected]>
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.
Stack from ghstack (oldest at bottom):
Add U55 and U85 16A8W quantization tests for linear operations following the pattern established in D80510628 for mul operations.
This adds test coverage for 16A8W (16-bit activations, 8-bit weights) quantization on both Ethos-U55 and Ethos-U85 platforms for linear operations. Both tests are marked with xfail annotations due to the known Vela compilation issue where int16 linear operations fail with "Conv2D Invalid arguments" error.
The tests follow the same structure as the existing 16A8W TOSA test but target specific hardware platforms:
Both tests use the same symmetric A16W8 quantization configuration and are properly decorated with platform-specific skip conditions (XfailIfNoCorstone300/320) and pytest xfail markers referencing GitHub issue 13947.
Addresses: #13729
Differential Revision: D82618039