strisunshinewentingwang/pytorch

Pull Requests Commits

[PyTorch] Optimize no input NVTX collection (#70133)

chaekit committed 2 years ago

64e30f6a

NNAPI: Add qint16 support via int16 (#70621)

Akshit Khurana committed 2 years ago

e8ce01cc

[PyTorch][Static Runtime] Fast aliasing in select_tensor by manual borrowing (#68122)

swolchok committed 2 years ago

eb9950aa

[PyTorch][Static Runtime] Support memory planning for torch.to() w/o requiring copying (#67223)

swolchok committed 2 years ago

a32450bd

[PyTorch][Static Runtime] Avoid managed output tensor DCHECK (#67221)

swolchok committed 2 years ago

5407963f

[PyTorch][Static Runtime] Fix dtype changing between iterations for to() (#67394)

swolchok committed 2 years ago

5b85f0dc

Support cloning CSR tensors (#70581)

pearu committed 2 years ago

006780b0

Change to use current LLLVM APIs (#70625)

modiking committed 2 years ago

8b5f1fe7

Avoid adding torch::deploy interpreter library to the data section (#70208)

PaliC committed 2 years ago

9421b3fd

Fix genSparseCSRTensor: generate non-trivial values for uint8 dtype. (#70580)

pearu committed 2 years ago

118d6800

[CUDA graphs] Changes batchnorm to increment num_batches_tracked in place for improved graph safety (#70444)

mcarilli committed 2 years ago

8faa1448

Remove backward op for slow 3d transposed convolution (#69933)

jbschlosser committed 2 years ago

58045720

Revert D33409880: [pytorch][PR] Deprecating Python 3.6

suo committed 2 years ago

09ef3ad9

[maskrcnn] use stable sort in mask rcnn caffe2 ops (#70510)

wat3rBro committed 2 years ago

1c8e7a6e

don't unsqueeze every stack arg if possible (#70288)

ngimel committed 2 years ago

5efda00c

Copy: Re-use existing neg and conj kernel implementations (#68949)

peterbell10 committed 2 years ago

206885d4

add SparseXPU to dispatch key set autogradother_backends (#70443)

majing921201 committed 2 years ago

f2004b5e

fractional_max_pool2d_backward: port to structured kernel (#68245)

francescocastelli committed 2 years ago

0965cea8

Sets device guard in _cudnn_impl functions (#70406)

ngimel committed 2 years ago

ef625642

Extend checkout for pytorch/builder (#70644)

janeyx99 committed 2 years ago

47853947

Revert "Revert D32498570: make codegen'd device guards not cuda-specific. Allow them to be used in external codegen" (#69951)

bdhirsh committed 2 years ago

b6ead419

Revert "Revert D32498569: allow external backend codegen to toggle whether to generate out= and inplace kernels" (#69950)

bdhirsh committed 2 years ago

1a82b6d4

Deprecating Python 3.6 (#70493)

atalman committed 2 years ago

9fcdec73

Flaky tests reporting: use GITHUB_RUN_ID instead of concatenated value (#70604)

janeyx99 committed 2 years ago

75692496

Revert D33198155: Bump version number to 7 and compile old operators with old schema

suo committed 2 years ago

04aab09d

Fix docstring for nn.ELU (#70574)

lithuak committed 2 years ago

cdcb7883

DispatchStub: Improve type mismatch errors (#67880)

peterbell10 committed 2 years ago

61283113

ibm z14/15 SIMD support (#66407)

quickwritereader committed 2 years ago

72f034cd

[DataPipe] Renaming FileLoader to FileOpener with deprecation warning for FileLoader (#70367)

NivekT committed 2 years ago

8505bc7e

Populate __name__ for torch.nn.modules.utils.{_single,_pair,...} (#70459)

jansel committed 2 years ago

7a3b039b

Newer Older