strisunshinewentingwang/pytorch

Pull Requests Commits

dbr quant: break up test class into multiple classes (#70246)

vkuzo committed 2 years ago

4e90fa6a

dbr quant: start recording ops which are not quantizeable (#70200)

vkuzo committed 2 years ago

5b200528

dbr quant: start reusing static quant module mappings (#70196)

vkuzo committed 2 years ago

80e685e2

Fix generating files for Vulkan on Windows (#69696)

mattjgalloway committed 2 years ago

45f5a3ce

[caffe2] don't use __FUNCSIG__ when building for Windows with clang (#70561)

h-friederich committed 2 years ago

c468e35d

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

chaekit committed 2 years ago

12653be4

NNAPI: Add qint16 support via int16 (#70621)

Akshit Khurana committed 2 years ago

44283c27

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

swolchok committed 2 years ago

10b40acb

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

swolchok committed 2 years ago

4d8fc869

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

swolchok committed 2 years ago

1507ce90

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

swolchok committed 2 years ago

99a10c37

Support cloning CSR tensors (#70581)

pearu committed 2 years ago

ab7d0df4

Change to use current LLLVM APIs (#70625)

modiking committed 2 years ago

d1dbcb17

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

PaliC committed 2 years ago

f8eaebc9

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

pearu committed 2 years ago

2292520b

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

mcarilli committed 2 years ago

29ff596d

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

jbschlosser committed 2 years ago

14457bb8

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

suo committed 2 years ago

1adb70c6

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

wat3rBro committed 2 years ago

8369a464

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

ngimel committed 2 years ago

b16b4448

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

peterbell10 committed 2 years ago

f8f96d48

add SparseXPU to dispatch key set autogradother_backends (#70443)

majing921201 committed 2 years ago

95a19526

fractional_max_pool2d_backward: port to structured kernel (#68245)

francescocastelli committed 2 years ago

a60adc7f

Sets device guard in _cudnn_impl functions (#70406)

ngimel committed 2 years ago

7e58b1dd

Extend checkout for pytorch/builder (#70644)

janeyx99 committed 2 years ago

6089a0f1

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

7b8c43cd

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

bdhirsh committed 2 years ago

bb5b4cce

Deprecating Python 3.6 (#70493)

atalman committed 2 years ago

d95be995

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

janeyx99 committed 2 years ago

4d08db0c

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

suo committed 2 years ago

0ece9a49

Newer Older