Skip to content

Pull requests: pytorch/ao

New pull request New
420 Open 3,388 Closed
420 Open 3,388 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Refactor da8w4 test to use shared model architecture CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4681 opened Aug 2, 2026 by abdulahi-banji Loading…
3
Fix autocast device handling for non-CUDA quantized linear paths
#4680 opened Aug 2, 2026 by saisharan0103 Loading…
1
2
Handle ROCm nightly gfx950 float8 sparse and quantization test limitations ciflow/rocm module: rocm
#4679 opened Aug 2, 2026 by saisharan0103 Loading…
3
Validate int4 tile-packed block_size against scale_and_zero
#4678 opened Aug 2, 2026 by saisharan0103 Loading…
1
2
Fix transposed fp8 row-major store mask axes
#4677 opened Aug 2, 2026 by saisharan0103 Loading…
1
2
Decouple int matmul op schemas from Triton availability
#4676 opened Aug 2, 2026 by saisharan0103 Loading…
2
Make MPS op imports side-effect safe during package discovery
#4675 opened Aug 2, 2026 by saisharan0103 Loading…
1
2
Validate quantization qparam dimensions before reshaping
#4674 opened Aug 2, 2026 by saisharan0103 Loading…
2
Document aarch64 regression checkout failure as transient CI infrastructure
#4673 opened Aug 2, 2026 by saisharan0103 Loading…
2
Allow GPTQ quantization on non-CUDA devices
#4672 opened Aug 1, 2026 by saisharan0103 Loading…
1
2
Fix inverted Linear bias flag in 1x1 Conv2d replacement
#4671 opened Aug 1, 2026 by saisharan0103 Loading…
1
2
Fix int_scaled_matmul tail masking for non-BLOCK_K K values
#4670 opened Aug 1, 2026 by saisharan0103 Loading…
2
[torchao] Add torch-free linear kernel target for ExecuTorch consumers CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4669 opened Aug 1, 2026 by digantdesai Contributor Loading…
1
Fold quantization into mutable buffer storage (#4665) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. meta-exported
#4665 opened Jul 31, 2026 by ethansfng Loading…
2
make tensor subclass copy_ error report exactly which metadata mismatched CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4664 opened Jul 30, 2026 by sainikhiljuluri Loading…
1
3
[test] Fix dead device comparison that disabled the SM80 bfloat16 skip in test/integration/test_integration.py ciflow/xpu label used to trigger xpu CI jobs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: not user facing Use this tag if you don't want this PR to show up in release notes xpu Intel XPU related features
#4660 opened Jul 30, 2026 by karol-brejna-i Loading…
3
Add core support for static Range Quantization (SRQ) and sub-byte embedding optimizations
#4659 opened Jul 30, 2026 by pculliton • Draft
2
[xpu][test][integration] Enable test_export_float8 on XPU ciflow/xpu label used to trigger xpu CI jobs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: not user facing Use this tag if you don't want this PR to show up in release notes xpu Intel XPU related features
#4658 opened Jul 30, 2026 by karol-brejna-i Loading…
3
Add _MXFP8GroupedMM variant with bf16 backward CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: training quantize_ api training flow
#4649 opened Jul 29, 2026 by andrewor14 Contributor Loading…
2
Mask scale loads in blockwise fp8 training GEMMs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4648 opened Jul 29, 2026 by guptaishaan Loading…
1
1
[torchao] Add torch-free linear kernel target for ExecuTorch consumers CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4645 opened Jul 29, 2026 by digantdesai Contributor Loading…
1
[xpu][test][dtypes] Finish XPU enablement of test_affine_quantized_float ciflow/xpu label used to trigger xpu CI jobs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: not user facing Use this tag if you don't want this PR to show up in release notes xpu Intel XPU related features
#4637 opened Jul 29, 2026 by karol-brejna-i Loading…
6
[xpu][test][float8] Enable test_float8_utils on XPU ciflow/xpu label used to trigger xpu CI jobs CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: not user facing Use this tag if you don't want this PR to show up in release notes xpu Intel XPU related features
#4636 opened Jul 29, 2026 by karol-brejna-i Loading…
5
Fix partial K tile masking in int_scaled_matmul CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4635 opened Jul 29, 2026 by bm1016bm-svg Loading…
1
3
Fix speedup column name for FP4 recipes in bench_matmul CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4632 opened Jul 28, 2026 by jagadish-amd Contributor Loading…
2
ProTip! Filter pull requests by the default branch with base:main.