strisunshinewentingwang/pytorch

Pull Requests Commits

Implement n-dimensional hermitian FFTs (#63890)

peterbell10 committed 2 years ago

6285348f

Add __module__ to torch.dtype.__dict__ (#65182)

nelhage committed 2 years ago

70f9f58a

[PyTorch][Edge] Fix inefficiency in objLoaderMobile (#65710)

swolchok committed 2 years ago

38c77539

[MicroBench] Added a micro benchmark for prefix sum (#65790)

navahgar committed 2 years ago

8f398325

[ci] fix softmax bc check (#65952)

suo committed 2 years ago

24f59fa2

Refactor `OperatorSupport` related code and fix TRT not supporting int64 dtype (#65848)

kflu committed 2 years ago

d4d3bb91

Revert D31238123: [pytorch][PR] Avoid saving self for`softmax` and `log_softmax`

suo committed 2 years ago

9ae63bd8

Add cuSPARSE descriptors and update CSR addmm (#60838)

IvanYashchuk committed 2 years ago

541eb1db

Update git version for CentOS base dockers (#65703)

jithunnair-amd committed 2 years ago

be00f020

[ci] try installing libgnutls to fix cert error (#65934)

suo committed 2 years ago

8297a16c

Move ASAN to GHA (#65846)

malfet committed 2 years ago

6a30d835

.github: Bump linux and windows gpu max available (#65923)

seemethere committed 2 years ago

cdbfb2b6

[JIT] Fix compilation unit reference link in constant object upon load (#65784)

Elias Ellison committed 2 years ago

928a4bba

[DataPipe] Fixes an issue where TarArchiveReader closes stream when read into a buffer (#65877)

NivekT committed 2 years ago

81301575

[RFC] [Modular] Include less headers in vararg_functions.cpp (#65672)

iseeyuan committed 2 years ago

7f87ff18

Update CODEOWNERS for optim (#65773)

albanD committed 2 years ago

ea776fa0

Convert Sampler back to lazily construction (#63646)

ejguan committed 2 years ago

b777d790

[quant] update fused_obs_fake_quant op to accept output_fake_quant argument (#65621)

supriyar committed 2 years ago

4666e3f1

[quant] adding memoryless observers for embeddingbag QAT work (#65699)

HDCharles committed 2 years ago

6d4b93bd

Revert D31132861: Make JIT Aliasing Test Less Brittle

suo committed 2 years ago

de80aff7

[Static Runtime] Disable SigridTransform + ListUnpack fusion when outputs reachable from graph output (#62697)

d1jang committed 2 years ago

4176afc4

[DatePipe] add deprecation warnings for DataPipes that will solely exist in TorchData (#65827)

NivekT committed 2 years ago

edab202a

[JIT] Make output of prim::TupleConstruct alias only with its inputs (#64879)

d1jang committed 2 years ago

cd458fe0

Skip failing tests when LAPACK and MAGMA are not available (#64930)

IvanYashchuk committed 2 years ago

dd354117

Remove "SciPioneer" from PT Distributed code owners (#65862)

Yi Wang committed 2 years ago

2c29ec2a

Revert D31005792: [NCCL] Init dummy NCCL comms in constructor

Mike Ruberry committed 2 years ago

91f8755b

Migrate THCIntegerDivider.cuh to ATen (#65745)

ngimel committed 2 years ago

5349ea92

(torchelastic) make --max_restarts explicit in the quickstart and runner docs (#65838)

Kiuk Chung committed 2 years ago

3900509b

[quant] Add imports to the torch/ao/quantization/__init__.py (#64911)

Zafar Takhirov committed 2 years ago

c7ef620a

Avoid saving self for`softmax` and `log_softmax` (#65242)

soulitzer committed 2 years ago

fb412bdd

Newer Older