Add alg_id argument to SemiSparseWeightConfig (#4238) by RandySheriff · Pull Request #4238 · pytorch/ao

RandySheriff · 2026-04-03T20:27:29Z

Summary:

As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~

pytorch-bot · 2026-04-03T20:27:34Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4238

📄 Preview Python docs built from this PR

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

✅ No Failures

As of commit c873eea with merge base d26bbae ():
💚 Looks good so far! There are no failures yet. 💚

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

meta-codesync · 2026-04-03T20:27:37Z

@RandySheriff has exported this pull request. If you are a Meta employee, you can view the originating Diff in D99485146.

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Differential Revision: D99485146

RandySheriff · 2026-04-03T22:11:18Z

@pytorchbot retest this please

pytorch-bot · 2026-04-03T22:12:44Z

❌ 🤖 pytorchbot command failed:

@pytorchbot: error: argument command: invalid choice: 'retest' (choose from 'merge', 'revert', 'rebase', 'label', 'drci', 'lint', 'fix-lint', 'apply-lint', 'cherry-pick')

usage: @pytorchbot [-h]
                   
                   {merge,revert,rebase,label,drci,lint,fix-lint,apply-lint,cherry-pick}
                   ...

Try @pytorchbot --help for more info.

RandySheriff · 2026-04-03T22:13:45Z

@pytorchbot fix-lint

test/core/test_config.py

jerryzh168 · 2026-04-03T22:26:19Z

@pytorchbot fix-lint

this doesn't work in torchao right now btw, I'm trying to figure out how to get something like this work

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Differential Revision: D99485146

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ https://www.internalfb.com/sandcastle/workflow/1247497096797230266 Differential Revision: D99485146

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Differential Revision: D99485146

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Fixes: - Fixed typo "activiation" → "activation" in test - Fixed formatting: line length violations and inconsistent indentation in `mm_search` call - Moved `__post_init__` API usage logging into `__init__` in `SemiSparseWeightConfig`, since the class is not a `dataclass` and `__post_init__` was dead code --- > Generated by [RACER](https://www.internalfb.com/wiki/RACER_(Risk-Aware_Code_Editing_and_Refactoring)/), powered by [Confucius](https://www.internalfb.com/wiki/Confucius/Analect/Shared_Analects/Confucius_Code_Assist_(CCA)/) [Session](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Chat), [Trace](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Trace) Differential Revision: D99485146

Summary: Pull Request resolved: pytorch#4238 As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Fixes: - Fixed typo "activiation" → "activation" in test - Fixed formatting: line length violations and inconsistent indentation in `mm_search` call - Moved `__post_init__` API usage logging into `__init__` in `SemiSparseWeightConfig`, since the class is not a `dataclass` and `__post_init__` was dead code --- > Generated by [RACER](https://www.internalfb.com/wiki/RACER_(Risk-Aware_Code_Editing_and_Refactoring)/), powered by [Confucius](https://www.internalfb.com/wiki/Confucius/Analect/Shared_Analects/Confucius_Code_Assist_(CCA)/) [Session](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Chat), [Trace](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Trace) Differential Revision: D99485146

Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Fixes: - Fixed typo "activiation" → "activation" in test - Fixed formatting: line length violations and inconsistent indentation in `mm_search` call - Moved `__post_init__` API usage logging into `__init__` in `SemiSparseWeightConfig`, since the class is not a `dataclass` and `__post_init__` was dead code --- > Generated by [RACER](https://www.internalfb.com/wiki/RACER_(Risk-Aware_Code_Editing_and_Refactoring)/), powered by [Confucius](https://www.internalfb.com/wiki/Confucius/Analect/Shared_Analects/Confucius_Code_Assist_(CCA)/) [Session](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Chat), [Trace](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Trace) Differential Revision: D99485146

…#4238) Summary: As titled, by adding the alg_id, we now have chance to select most appropriate algorithm for optimal perf for specific gemm shapes ~ Fixes: - Fixed typo "activiation" → "activation" in test - Fixed formatting: line length violations and inconsistent indentation in `mm_search` call - Moved `__post_init__` API usage logging into `__init__` in `SemiSparseWeightConfig`, since the class is not a `dataclass` and `__post_init__` was dead code --- > Generated by [RACER](https://www.internalfb.com/wiki/RACER_(Risk-Aware_Code_Editing_and_Refactoring)/), powered by [Confucius](https://www.internalfb.com/wiki/Confucius/Analect/Shared_Analects/Confucius_Code_Assist_(CCA)/) [Session](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Chat), [Trace](https://www.internalfb.com/confucius?session_id=b3d5061c-31f8-11f1-9a22-a5a2fb0b80aa&tab=Trace) Differential Revision: D99485146

RandySheriff requested a review from jerryzh168 as a code owner April 3, 2026 20:27

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 Apr 3, 2026

meta-codesync bot added fb-exported meta-exported labels Apr 3, 2026

RandySheriff added the module: inference quantize_ api inference flow label Apr 3, 2026

meta-codesync bot changed the title ~~Add alg_id argument to SemiSparseWeightConfig~~ Add alg_id argument to SemiSparseWeightConfig (#4238) Apr 3, 2026

RandySheriff force-pushed the export-D99485146 branch from 5f77bd0 to 72c9568 Compare April 3, 2026 21:45

RandySheriff requested a review from vkuzo as a code owner April 3, 2026 21:45

RandySheriff force-pushed the export-D99485146 branch from 72c9568 to 3bd2dd7 Compare April 3, 2026 21:47

jerryzh168 reviewed Apr 3, 2026

View reviewed changes

test/core/test_config.py Outdated Show resolved Hide resolved

RandySheriff force-pushed the export-D99485146 branch from 3bd2dd7 to 8dd8c54 Compare April 3, 2026 23:33

RandySheriff force-pushed the export-D99485146 branch from 8dd8c54 to 4762088 Compare April 3, 2026 23:40

RandySheriff force-pushed the export-D99485146 branch from 4762088 to 5655c53 Compare April 6, 2026 19:17

RandySheriff force-pushed the export-D99485146 branch from 5655c53 to 9adab04 Compare April 6, 2026 19:20

RandySheriff force-pushed the export-D99485146 branch from 9adab04 to 00c66c9 Compare April 6, 2026 20:34

RandySheriff force-pushed the export-D99485146 branch from 00c66c9 to bac71e1 Compare April 6, 2026 20:49

RandySheriff force-pushed the export-D99485146 branch 2 times, most recently from ebd1465 to 0ba7e5a Compare April 6, 2026 20:53

RandySheriff force-pushed the export-D99485146 branch from 0ba7e5a to 7e407ab Compare April 6, 2026 21:05

RandySheriff force-pushed the export-D99485146 branch from 7e407ab to b78b831 Compare April 6, 2026 21:17

RandySheriff force-pushed the export-D99485146 branch from b78b831 to c6e63f5 Compare April 6, 2026 22:35

RandySheriff force-pushed the export-D99485146 branch from c6e63f5 to 16bc4a8 Compare April 6, 2026 22:41

RandySheriff force-pushed the export-D99485146 branch from 16bc4a8 to be72211 Compare April 6, 2026 22:49

jerryzh168 approved these changes Apr 6, 2026

View reviewed changes

RandySheriff force-pushed the export-D99485146 branch from be72211 to d8f20dc Compare April 7, 2026 00:27

RandySheriff force-pushed the export-D99485146 branch from d8f20dc to 01eb4ad Compare April 7, 2026 02:47

RandySheriff force-pushed the export-D99485146 branch from 01eb4ad to d97a6ec Compare April 7, 2026 05:34

RandySheriff force-pushed the export-D99485146 branch from d97a6ec to c873eea Compare April 7, 2026 05:37

jerryzh168 merged commit 2a8fa55 into pytorch:main Apr 7, 2026
21 checks passed

Freed-Wu mentioned this pull request Apr 12, 2026

Add torch.uint16, torch.uint32 #4269

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add alg_id argument to SemiSparseWeightConfig (#4238)#4238

Add alg_id argument to SemiSparseWeightConfig (#4238)#4238
jerryzh168 merged 1 commit intopytorch:mainfrom
RandySheriff:export-D99485146

RandySheriff commented Apr 3, 2026 •

edited

Loading

Uh oh!

pytorch-bot bot commented Apr 3, 2026 •

edited

Loading

Uh oh!

meta-codesync bot commented Apr 3, 2026

Uh oh!

RandySheriff commented Apr 3, 2026

Uh oh!

pytorch-bot bot commented Apr 3, 2026

Uh oh!

RandySheriff commented Apr 3, 2026

Uh oh!

Uh oh!

jerryzh168 commented Apr 3, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

RandySheriff commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Apr 3, 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/ao/4238

✅ No Failures

Uh oh!

meta-codesync bot commented Apr 3, 2026

Uh oh!

RandySheriff commented Apr 3, 2026

Uh oh!

pytorch-bot bot commented Apr 3, 2026

Uh oh!

RandySheriff commented Apr 3, 2026

Uh oh!

Uh oh!

jerryzh168 commented Apr 3, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

RandySheriff commented Apr 3, 2026 •

edited

Loading

pytorch-bot bot commented Apr 3, 2026 •

edited

Loading