strisunshinewentingwang/pytorch

Pull Requests Commits

CUDACachingAllocator: Keep one event queue per stream (#71745)

nelhage committed 2 years ago

c585d354

Fix upgrader codegen when constant list is 0 (#72199)

cccclai committed 2 years ago

d23231fd

Fix torch.save and detach for CSR Tensor (#71963)

cpuhrsch committed 2 years ago

4a7e07e5

ci: Migrate macOS x86_64 binary builds to GHA (#71888)

seemethere committed 2 years ago

c2e63b43

Improve fx2trt benchmark (#72145)

Yinghai Lu committed 2 years ago

4c62ffa1

Fix unused variable warning in LostCTC.cu (#72155)

r-barnes committed 2 years ago

1ad53b51

Add hook for functorch to error out with unoverridable autograd operations (#72176)

zou3519 committed 2 years ago

774e0847

Set correct device id on efficientzerotensors (#71611)

anjali411 committed 2 years ago

f607af12

Update approvers for ONNX (#71659)

garymm committed 2 years ago

889a62dd

[SR] Implement prim::Loop (#69838)

Mike Iovine committed 2 years ago

2d5296b0

[SR] Implement prim::If (#69837)

Mike Iovine committed 2 years ago

2aa69950

[SR] Force sub-blocks to return at least one output (#69836)

Mike Iovine committed 2 years ago

d2599701

[SR] Graph pass to create owned refs of special IValues (#69835)

Mike Iovine committed 2 years ago

238dded1

Codegen: Do less work in dry-runs for sharded files (#69805)

peterbell10 committed 2 years ago

b0518b27

[Profiler] Defer KinetoEvent and GenericTraceActivity creation to post processing. (#71539)

Taylor Robie committed 2 years ago

cb0d7f0d

enable GHA workflow defaults for ROCm (#72142)

jeffdaily committed 2 years ago

e19f2e52

fix typos in aten/src/ATen/native/mkldnn (#71853)

jinyouzhi committed 2 years ago

31b34841

[fix] Dropout2d-3d no-batch-dim (#69885)

kshitij12345 committed 2 years ago

02f6226b

Updated CONTRIBUTING.md (#72044)

NeoKish committed 2 years ago

39483b59

Reland torch.ops API change machinery with the core functionality disabled (#71785)

anjali411 committed 2 years ago

a1383a9c

Make `asarray` behavior consistent with Python Array API. (#71757)

ysiraichi committed 2 years ago

1fdbe9aa

make fsdp folder to be public (#72084)

zhaojuanmao committed 2 years ago

2336571c

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

ed435e90

Update logspace and bump the version number to 9 (#72051)

tugsbayasgalan committed 2 years ago

8757e21c

Jiterator cache for Windows (#72048)

Sacha Refshauge committed 2 years ago

230186f9

[reland] Create torch.distributed._shard package. (#72141)

pritamdamania committed 2 years ago

64670e41

[DataPipe] Disable Typing for DataPipe before branch cut (#72123)

ejguan committed 2 years ago

7b014cc6

Check for grad mode enabled in dynamic shape fusion check (#72161)

Elias Ellison committed 2 years ago

aa99df5c

change default grain size in gather scatter kernel to improve CPU performance (#72082)

mingfeima committed 2 years ago

0c49800c

Fix fx docs (#72108)

mrshenli committed 2 years ago

7c2eda38

Newer Older