strisunshinewentingwang/pytorch

Pull Requests Commits

Back out "[pytorch][PR] Merge CUDA Streams and Events" (#54996)

jianyuh committed 3 years ago

7fc03dd7

[FX] Added FX prepare_for_inference for Intel CPUs (#53805)

Chillee committed 3 years ago

24bfcd53

[Pytorch Mobile] 'fix' filter of named parameters for FL (#54633)

JacobSzwejbka committed 3 years ago

a0ae3e52

[ROCm] add 4.1 docker image (#54628)

jeffdaily committed 3 years ago

0dd873bd

cmake: fix ONNX_NAMESPACE if USE_SYSTEM_ONNX (#54973)

SpaceIm committed 3 years ago

aeedd5c7

Update Kineto submodule (#55011)

Ilia Cherniavskii committed 3 years ago

449a9514

Automated submodule update: tensorpipe (#54970)

facebook-github-bot committed 3 years ago

99a423f8

Revert D27370295: [android] fbjni android use prefab dependency, version 0.2.2

agolynski committed 3 years ago

09756e72

Revert D27422219: [caffe2] Support Log1p operator

agolynski committed 3 years ago

25e07c6e

Added support for TensorList inputs in OpInfo (#54922)

heitorschueroff committed 3 years ago

6d87b366

[package] fix mangling issues with TorchScript (#54915)

suo committed 3 years ago

8a170fba

Add Type System (I) (#53244)

nikithamalgi committed 3 years ago

444e5f0b

[PyTorch] Add DimVector variant of infer_size (#54882)

swolchok committed 3 years ago

4865195c

[android] fbjni android use prefab dependency, version 0.2.2 (#54792)

IvanKobzarev committed 3 years ago

2bee09a5

Fixed the default size of the workspace array for MAGMA's SVD (#54875)

IvanYashchuk committed 3 years ago

854c9207

[ROCm] utilize PUBLIC vs PRIVATE linking to avoid incorrect dependencies (#54727)

jeffdaily committed 3 years ago

1dffbe75

Mark redispatch functions with TORCH_API (#54966)

nunoplopes committed 3 years ago

d4c37b31

[ROCm] skip some tests to enable 4.1 CI upgrade (#54536)

arindamroy-eng committed 3 years ago

b907d6e3

Added Azure Pipelines build steps for PyTorch (#54039)

mstfbl committed 3 years ago

3baeeb3f

lazily init AliasDb and add `changed` status to CSE (#54776)

Krovatkin committed 3 years ago

f956b752

Remove hacky wrapper for about 70 kernels (#54898)

wenleix committed 3 years ago

2df4acd0

Remove hacky wrappers for 21 operators (#54819)

wenleix committed 3 years ago

1bf57430

[caffe2] Support Log1p operator (#54968)

khabinov committed 3 years ago

d92e2520

[PyTorch] One less refcount bump in linear() (#54936)

swolchok committed 3 years ago

d490e012

[PyTorch] Avoid refcount bumps in addmm_out_cuda_impl (#54935)

swolchok committed 3 years ago

dde7fff0

[PyTorch] Avoid refcount bump in TensorArg (#54934)

swolchok committed 3 years ago

ea37fe34

Revert D25966661: Support needsOutputs for RecordFunction and ObserverUtil improvements

Qi Zhao committed 3 years ago

5b448cf2

test_c10d: use with_nccl_blocking_wait decorator (#54742)

rohan-varma committed 3 years ago

23b15ef9

test_c10d: Run tests with nccl_async_error_handling (#54741)

rohan-varma committed 3 years ago

3f1cd2e3

Provide a decorator to set/unset nccl blocking wait for tests (#54740)

rohan-varma committed 3 years ago

0e543b2b

Newer Older