Skip to content

Pull requests: pytorch/ao

New pull request New
327 Open 3,322 Closed
327 Open 3,322 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

support cat, chunk and split for MXTensor CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4485 opened Jun 13, 2026 by discobot Loading…
1
1
[pt2e] Fuse and fold linear + bn in QAT like conv + bn CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4484 opened Jun 13, 2026 by discobot Loading…
1
1
pt2e: raise clear errors for unresolvable SharedQuantizationSpec annotations (#4421) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4483 opened Jun 13, 2026 by discobot Loading…
1
1
[PT2E][X86] Extend reused-module fallback to nn.Conv2d (follow-up to #4480) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4482 opened Jun 13, 2026 by waynehacking8 Loading…
1
[PT2E][X86] Fall back to plain linear annotation for reused nn.Linear with fusable post op 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
#4480 opened Jun 12, 2026 by waynehacking8 Loading…
1
12
[Draft] [XPU] Add NVFP4 Support on Intel GPU CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4477 opened Jun 12, 2026 by xiaowangintel Collaborator Loading…
1
rowwise_scaled_linear_sparse: shape-based tile/cluster/schedule selection CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. meta-exported
#4476 opened Jun 12, 2026 by ZihaoLiu0927 Contributor Loading…
2
Enable pt2e XPU tests in test_x86inductor_fusion
#4474 opened Jun 11, 2026 by chunhuanMeng Loading…
2
Integrate DeepGEMM FP8 blockwise MoE backend 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
#4473 opened Jun 9, 2026 by iamzainhuda Contributor • Draft
1
[mxfp8] Skip weight gradient for frozen weights in mx_mm 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
#4471 opened Jun 9, 2026 by ultism Contributor Loading…
3
[moe training] Use non-atomic Triton scatter for permute_and_pad backward CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: rocm module: training quantize_ api training flow
#4469 opened Jun 8, 2026 by rishisinhanj Contributor Loading…
2
[Draft] [XPU] Add MXTensor Support on Intel GPU CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4468 opened Jun 8, 2026 by xiaowangintel Collaborator Loading…
1
Add workflow to triage main CI failures with Claude 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
#4467 opened Jun 5, 2026 by andrewor14 Contributor Loading…
11
Add FP8 Row-wise QAT for MoE CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4464 opened Jun 4, 2026 by CaoSY • Draft
3
fix(Int8DynActInt4WeightQuantizer): use scales_precision for runtime modules 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
#4459 opened Jun 2, 2026 by Anai-Guo Contributor Loading…
1 task
1
7
docs: clarify Float8DynamicActivationFloat8WeightConfig is inference-only CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4454 opened May 29, 2026 by swapnil5053 Loading…
1
1
fix: float8_transpose uses == instead of = when remapping _axiswise_dim CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4453 opened May 29, 2026 by swapnil5053 Loading…
1
1
fix: parametrized dtype ignored in test_autotuner; wrong variable in test_integration error message CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4452 opened May 29, 2026 by swapnil5053 Loading…
1
1
docs(Float8DynamicActivationFloat8WeightConfig): mark as inference-only (#4376) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4451 opened May 29, 2026 by Anai-Guo Contributor Loading…
1
1
[ROCm] Add HIP device synchronize event to profiler overhead filter ciflow/rocm CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: rocm
#4450 opened May 29, 2026 by jagadish-amd Contributor Loading…
6
bump version to 0.19.0 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
#4448 opened May 28, 2026 by vkuzo Contributor Loading…
1
Binary changes CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4439 opened May 26, 2026 by Priyjain-amd • Draft
1
fix(pt2e/constant_fold): preserve shared get_attr target when only some uses are dead (#4420) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4431 opened May 20, 2026 by Anai-Guo Contributor Loading…
1 of 2 tasks
1
4
Add _grouped_mm dispatch handler to MXTensor CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. module: inference quantize_ api inference flow moe mx topic: new feature Use this tag if this PR adds a new feature
#4419 opened May 18, 2026 by xiaowangintel Collaborator Loading…
12
Fix PT2E reference rewrite int kernel dtype mismatch on nightly CPU CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4417 opened May 17, 2026 by Copilot AI • Draft
@xiaowangintel
1
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.