strisunshinewentingwang/pytorch

Pull Requests Commits

test cupti-so branch

ezyang committed 2 years ago

1f6c1ea7

[Quant] Qlinear Add qint8 support backed by xnnpack (#73672)

digantdesai committed 2 years ago

d8dbf9d8

[Quant] Qconv: Add qint8 support backed by xnnpack (#73669)

digantdesai committed 2 years ago

bac6b8c8

[fx/operator_schemas] Bring back check for OpOverload (#73978)

jfix71 committed 2 years ago

b0a327d1

[GHF] Specify multiple mandatory checks

malfet committed 2 years ago

f71eede8

[tensorexp] ExternalCall2 without memcpy (#72225)

IvanKobzarev committed 2 years ago

519e226b

[Quant][test] Added test to check if fp16 packing->unpacking yields the same result as to(torch.float16).to(torch.float32) [reland] (#73808)

dzdang committed 2 years ago

bc512253

Fix deadlock in some edge case in autograd (#73961)

albanD committed 2 years ago

b2a55076

RNN args renaming in memonger.

xkszltl committed 2 years ago

adae0d35

ci: Fix cudatoolkit issue, make docker builds testable

seemethere committed 2 years ago

48024498

[GHF][BE] Add match_rules test

malfet committed 2 years ago

aed71885

Revert "Check that PR base is default branch in trymerge.py"

pytorchmergebot committed 2 years ago

63a1e7bd

[OSS] add script to generate test models for mobile (#73746)

linbinyu committed 2 years ago

69cdcd2d

[Lazy][JIT] Do not crash when target device is unsupported by fuser (#73820)

comaniac committed 2 years ago

012829eb

Get rid of TorchScript sparse tensor is experimental warning. (#73874)

ezyang committed 2 years ago

2f90c822

Relax dtype restrictions on torch.Tensor (#73850)

ezyang committed 2 years ago

02392843

[DataPipe] Slight refactoring IterDataPipe serialization test (#73922)

NivekT committed 2 years ago

8811d217

[DataPipe] Adding serialization test for all MapDataPipe (#73921)

NivekT committed 2 years ago

08211540

Reset worker cycle iterator for determinism across runs (#73675)

evren-cerebras committed 2 years ago

75345257

Print system info as part of EC2 info step

malfet committed 2 years ago

3c82e422

[ONNX] Add symbolic support for torch.nn.cosinesimilarity (#72128) (#73283)

BowenBao committed 2 years ago

97ae431e

[ONNX] use onnxruntime 1.10 in CI (#69271) (#73282)

BowenBao committed 2 years ago

95b12327

[ONNX] Add module name as PythonOp attribute (#67193) (#73281)

BowenBao committed 2 years ago

341e20a1

[ONNX] Adds overload_name to Aten op (#69378) (#73280)

BowenBao committed 2 years ago

9210e8f5

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

CodemodService FBSourceClangFormatLinterBot committed 2 years ago

343a9733

add script to aggregate production ops

linbinyu committed 2 years ago

a267c8e7

FX graph module - prevent infinite recursion (#73866)

anijain2305 committed 2 years ago

7ebab924

[PyTorch] Move is_nested_tensor_impl & add get_nested_tensor_impl_or_null (#73928)

swolchok committed 2 years ago

b9c59ddc

[pytorch] flatten_indices function should use vector::resize instead of reserve (#73831)

jiyuanzFB committed 2 years ago

ad6290f4

[SR][easy] Stack/concat out variants do not segfault on empty inputs (#73704)

Mike Iovine committed 2 years ago

97b20b9b

Older