-
Notifications
You must be signed in to change notification settings - Fork 455
Pull requests: pytorch/ao
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
soften version guard check for low precision attention API
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
#4058
opened Mar 11, 2026 by
howardzhang-cv
Loading…
soften version guard check for low precision attention API
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4057
opened Mar 11, 2026 by
howardzhang-cv
Loading…
Added prototype low precision attention API to the docs
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
topic: documentation
Use this tag if this PR adds or improves documentation
#4056
opened Mar 11, 2026 by
howardzhang-cv
Loading…
remove rope fusion option, do automatically on torch compile
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
#4055
opened Mar 11, 2026 by
howardzhang-cv
Loading…
[xpu][test] Skip WIP config for Intel GPU in test/prototype/safetensors/test_safetensors_support.py
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4049
opened Mar 11, 2026 by
zxd1997066
Loading…
Buckify the ao arm shared embedding kernels
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
module: not user facing
Use this tag if you don't want this PR to show up in release notes
#4047
opened Mar 11, 2026 by
JacobSzwejbka
Loading…
[xpu][test] Port test/prototype/quant_logger/test_quant_logger.py to Intel GPU
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4046
opened Mar 11, 2026 by
zxd1997066
Loading…
[ROCm] Enable inference quantization tests on ROCm (Float8, Int8, per-token)
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
module: rocm
#4044
opened Mar 11, 2026 by
brucechanglongxu
Loading…
Fix deprecated TorchAO dtype imports
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4043
opened Mar 11, 2026 by
drivanov
Loading…
[ROCm] Add NVFP4 emulation support for ROCm (MI300/MI350)
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
module: rocm
#4042
opened Mar 11, 2026 by
brucechanglongxu
Loading…
[ROCm] Enable MXFP8/MXFP4 emulation tests on ROCm (MI300+)
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
[CPU] Add check of AVX512 at runtitme
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
Fix axiswise dim assignment in float8_transpose
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4038
opened Mar 10, 2026 by
GotFusion
Loading…
[X86] Remove deprecated Int8DynamicActInt4WeightCPULayout
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
#4037
opened Mar 10, 2026 by
Xia-Weiwen
Loading…
Enable blockwise FP8 dense training kernels on ROCm
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
float8
module: rocm
chore: fix syntax warning
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
#4035
opened Mar 10, 2026 by
DefTruth
Loading…
[ROCm] Add structured device capability helpers for AMD GPUs
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
module: rocm
#4034
opened Mar 10, 2026 by
brucechanglongxu
Loading…
[X86] Re-enable DA8W4 path on X86 CPU
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
#4033
opened Mar 10, 2026 by
Xia-Weiwen
•
Draft
hook up mslk's to_nvfp4 kernel to torchao's inference nvfp4 workflows
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
#4031
opened Mar 9, 2026 by
vkuzo
Loading…
Fix Android ARM64 build for torchao lowbit kernels
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
#4029
opened Mar 9, 2026 by
SS-JIA
Loading…
remove 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
unwrap_tensor_subclass from tutorial
CLA Signed
#4028
opened Mar 9, 2026 by
jcaip
Loading…
Upgrade torchao build configs to C++20
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: for developers
Use this tag if this PR is mainly developer facing
#4025
opened Mar 8, 2026 by
r-barnes
Loading…
delete v1 of Int8WeightOnlyConfig
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
#4020
opened Mar 6, 2026 by
vkuzo
Loading…
delete v1 of Int8DynamicActivationInt8WeightConfig
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
#4019
opened Mar 6, 2026 by
vkuzo
Loading…
implement per-group quantization for Int8WeightOnlyConfig
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
#4018
opened Mar 6, 2026 by
vkuzo
Loading…
Previous Next
ProTip!
Type g p on any issue or pull request to go back to the pull request listing page.