strisunshinewentingwang/pytorch

Pull Requests Commits

onnx export of per channel fake quantize functions (#42835)

skyw committed 3 years ago

1cf3ff66

[cmake] Add explicit cublas->cudart dependency (#52243) (#52404)

malfet committed 3 years ago

Verified f7c4afc0

[1.8] Remove torch.vmap (#52397)

zou3519 committed 3 years ago

Verified 20554c00

[1.8] Fix libnvrtc discoverability in package patched by `auditwheel` (#52365)

malfet committed 3 years ago

Verified 3464d64f

Skip OneDNN Convolution in case of groups = 24 #50042 (#52313)

VitalyFedyunin committed 3 years ago

Verified c6972eb3

Use side-stream in CPU to GPU copies in DDP (#50180) (#52270)

rohan-varma committed 3 years ago

Verified 25562d3d

ports fix (#52242)

Mike Ruberry committed 3 years ago

Verified cd63c37b

[v1.8 patch] [Resubmission] Add a documentation page for DDP communication hooks (#52215)

Yi Wang committed 3 years ago

Verified c79decdb

[1.8] Do not print warning if CUDA driver not found (#51806) (#52050)

malfet committed 3 years ago

Verified c307a3f3

Workaround arm64 gcc error in `std::copysign` (#51900) (#52049)

malfet committed 3 years ago

Verified f0710207

fake_quant cachemask: remove Python bindings (#51878) (#52160)

vkuzo committed 3 years ago

Verified 4f436f85

[FX][1.8] Cherrypick three FX fixes to 1.8 (#52021)

James Reed committed 3 years ago

Verified ae115897

1.8 cherrypick: Add metacompile of Ternary if (#51789) (#51913)

ppwwyyxx committed 3 years ago

Verified 9e5bcc10

.jenkins: Release branch specific updates (#51982)

seemethere committed 3 years ago

Verified fa857824

[v1.8.0] [wip] doc_fix (#52006)

seemethere committed 3 years ago

Verified 13688095

[FX] Hide experimental folder (#51987)

James Reed committed 3 years ago

Verified 4073248f

Disable unaliged-access test from TestVectorizedMemoryAccess.CopyKernel (#51864) (#51890)

janeyx99 committed 3 years ago

Verified 75153cb7

concantenate LICENSE files when building a wheel (#51634) (#51882)

Rong Rong committed 3 years ago

Verified 5bb69b08

[FX][docs] Indent forward (#51802)

James Reed committed 3 years ago

9112f4ed

pytorch docs: add fake_quantize functions documentation (#51748)

vkuzo committed 3 years ago

8c48af82

[Conda][Kineto] efine weak `acc_get_device_type` if kineto is used (#51818)

malfet committed 3 years ago

ececbcff

Adding support for CUDA 11.2 in our nightly build matrix (#51611)

janeyx99 committed 3 years ago

fb07aca7

Add FLOPS support to the new profiler API. (#51734)

xuzhao9 committed 3 years ago

5c3a054b

Revert D26009829: Optimize relu on cpu using clamp_min

albanD committed 3 years ago

430329e8

Enable GPU/RE tags for caffe2/caffe2/python/TARGETS

Rong Rong (AI Infra) committed 3 years ago

50c9c082

Optimize relu on cpu using clamp_min (#50924)

bertmaher committed 3 years ago

2054cd56

[quick-checks] Allow `gradlew` to be executable (#51796)

malfet committed 3 years ago

3cfbf6d3

[Metal] Add hardswish and hardsigmoid to metal, fix broadcasting for binary elementwise ops

SS-JIA committed 3 years ago

029f857b

Revert D26276903: [pytorch][PR] Add LazyBatchNormXd

albanD committed 3 years ago

a930162c

Add `acc_get_device_type` weak symbol to `kineto_profler` (#51787)

malfet committed 3 years ago

33973d45

Older