Skip to content

Pull requests: pytorch/ao

New pull request New
245 Open 2,918 Closed
245 Open 2,918 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

Add Static Activation Quantization Subclass for Observation 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
#3925 opened Feb 20, 2026 by namgyu-youn Loading…
4 tasks
2
[mxfp8 moe training] migrate ep utils to mx_formats 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 moe mx topic: improvement Use this tag if this PR is an improvement (doesn't fit into any of the other categories)
#3922 opened Feb 20, 2026 by danielvegamyhre Loading…
1
Add MXFP4 support for GPTQ quantization and separate eval script CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3921 opened Feb 19, 2026 by jcaip Loading…
1
Lowbit Embedding on x86 CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported meta-exported
#3919 opened Feb 19, 2026 by JacobSzwejbka Loading…
2
Add Linear+BN folding to _fuse_conv_bn_() CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported meta-exported
#3917 opened Feb 19, 2026 by SS-JIA Loading…
2
Add support for flashinfer quantize kernel option for nvfp4 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
#3912 opened Feb 17, 2026 by jerryzh168 Loading…
7
Refactor use_triton_kernel to use nvfp4_quantize_kernel_choice 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
#3911 opened Feb 17, 2026 by jerryzh168 Loading…
20
Add parameter_name support to _intx_weight_only_transform 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
#3905 opened Feb 17, 2026 by jcaip Loading…
1
Add parameter_name support to _int8_dynamic_activation_intx_weight_transform 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
#3904 opened Feb 17, 2026 by jcaip Loading…
1
Fix _int8_dynamic_activation_int8_weight_transform to use parameter_name properly 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
#3903 opened Feb 17, 2026 by jcaip Loading…
1
Add parameter_name support to _float8_dynamic_activation_int4_weight_transform 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
#3902 opened Feb 17, 2026 by jcaip Loading…
1
Add parameter_name support to _int4_weight_only_transform 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
#3901 opened Feb 17, 2026 by jcaip Loading…
2
Add asymmetric support for Int8Tensor + SmoothQuant 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
#3900 opened Feb 17, 2026 by jcaip Loading…
1
refactor GPTQ observer for dynamic quant CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3899 opened Feb 17, 2026 by jcaip Loading…
1
[mxfp8 moe training] move mxfp8 grouped mm code into mx_formats 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 moe mx topic: improvement Use this tag if this PR is an improvement (doesn't fit into any of the other categories)
#3898 opened Feb 17, 2026 by danielvegamyhre Loading…
4
Add MXFP8 support for GPTQ quantization CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3897 opened Feb 17, 2026 by jcaip Loading…
1
Add precomputed scale support to MXTensor and fix swizzled scales 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
#3895 opened Feb 17, 2026 by jcaip Loading…
2
WIP [mx][test] Refactor skip logic in test_mx_linear CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3872 opened Feb 12, 2026 by ugolowic • Draft
6
[QAT] Add low-bit FSDP all-gather path for affine fake quant tensor CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3871 opened Feb 12, 2026 by MagellaX Loading…
1
2
Added benchmarking for new torchao low precision attention api benchmark 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
#3865 opened Feb 12, 2026 by howardzhang-cv Loading…
1
Added new API for low precision fp8 attention using FA3 CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. topic: new feature Use this tag if this PR adds a new feature
#3857 opened Feb 11, 2026 by howardzhang-cv Loading…
8
Add option to skip scaling with global scaling factor for nvfp4 in torchao CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported meta-exported topic: improvement Use this tag if this PR is an improvement (doesn't fit into any of the other categories)
#3851 opened Feb 10, 2026 by jbschlosser Loading…
2
Add mixed granularity tests for smooth quant CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3847 opened Feb 10, 2026 by cyxlily • Draft
1
[ROCm][INT4] Corner case n<Ntile handling ciflow/rocm CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. device: rocm
#3834 opened Feb 6, 2026 by ZhiweiYan-96 Loading…
6
Test with Helion instead of Triton for fp8 rope sdpa CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#3830 opened Feb 5, 2026 by howardzhang-cv • Draft
1
ProTip! Updated in the last three days: updated:>2026-02-17.