strisunshinewentingwang/pytorch

Pull Requests Commits

Raise TypeError when the argument to isinf and isfinite is not a tensor (#20817)

xuhdev committed 5 years ago

ef1fdc27

Fix documentation for attention mask shape (#20850)

jotsif committed 5 years ago

87040af4

Use "length of the RNN input" instead of "length of the RNN"

kit1980 committed 5 years ago

a5c90aaf

Instructions for how to update pytorch-ci-hud when updating binary builds (#20758)

ezyang committed 5 years ago

3e4f213e

Resend "Split ATen/Parallel into interface and backend" (#20825)

Ilia Cherniavskii committed 5 years ago

c3d05e86

Fix init_thread calls in thread pool initialization (#20848)

Ilia Cherniavskii committed 5 years ago

6b748567

Change the quantizer to match the behavior of the FBGEMM implementation (#20892)

Zafar Takhirov committed 5 years ago

1bb728fe

Catchall kernels instead of fallback kernels (#20773)

smessmer committed 5 years ago

fc941d3b

Lightweight at-most-once logging for API usage (#20745)

Dmytro Dzhulgakov committed 5 years ago

c25e3378

Remove Variable::Impl and DifferentiableViewImpl (#17072)

Will Feng committed 5 years ago

8cde4c4d

Adding ShufflenetV2 to caffe2's benchmark suite. (#20180)

Abhinav Jauhri committed 5 years ago

f93e0619

Updating submodules

svcscm committed 5 years ago

3aa7ee6f

Updating submodules

svcscm committed 5 years ago

cfb6c4a8

dropout symbolic_script should respect the training flag (#20760)

suo committed 5 years ago

62af37aa

Move torchvision install out of onnx test script

bddppq committed 5 years ago

bd53c8eb

Dict is a reference type (#20669)

smessmer committed 5 years ago

d5b7138a

bug fix 19374 - fix for upsample export

pk-g committed 5 years ago

93d5503f

Fix oscillation in coalesceInsertedDataDependencies (#20833)

Yinghai Lu committed 5 years ago

48bf7b9b

Use conda torchvision version (#20865)

Will Feng committed 5 years ago

2d96876d

Move THCTensor_{random, clampedRandom, cappedRandom} to ATen (#20620)

syed-ahmed committed 5 years ago

b6d0f6c8

Avoid dynamic dispatch inside the omp loop in AdaptiveAvgPool2d (#20366)

crcrpar committed 5 years ago

48424a6c

Modify cos to cosh in Vec256 (#20797)

vishwakftw committed 5 years ago

cf0268e5

Add mkldnn sigmoid operator

bddppq committed 5 years ago

70caa2ef

Enable torch.jit.trace for mkldnn modules

bddppq committed 5 years ago

8dedb04c

Add support for save and load mkldnn modules

bddppq committed 5 years ago

63585c3b

Fix build error with MSVC (#20853)

Jonas1312 committed 5 years ago

5f83c5d8

Dictionarize check_inputs coming from `trace`

Krovatkin committed 5 years ago

31e2d20c

fix the input/output type mismatch (#20829)

lly-zero-one committed 5 years ago

2c556a94

Make LayerNorm.normalized_shape a tuple

davidriazati committed 5 years ago

9c57d8df

Fixes error with custom scalars, fixes #20579 (#20580)

baldassarreFe committed 5 years ago

99b3f5cd

Newer Older