strisunshinewentingwang/pytorch

Pull Requests Commits

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

Workaround python2.7 find_module limitation / explicitly close file (#20782)

Ailing Zhang committed 5 years ago

a16708a1

Port dilated_max_pool2d() to ATen

skrah committed 5 years ago

ec57d1f1

Add back at::_copy_from for use by XLA (#20783)

colesbury committed 5 years ago

f039401b

fix a couple of typos in README markdown (#20819)

nairbv committed 5 years ago

80aed36f

add batch of string ops (#20826)

Elias Ellison committed 5 years ago

8fc069fa

Install torchvision from master

bddppq committed 5 years ago

90182a73

Remove cpu_half, cpu_bool, cuda_bool from native_functions.yaml (#20552)

Roy Li committed 5 years ago

d35a5879

Add PerChannelAffineQuantizer (#20764)

jerryzh168 committed 5 years ago

41100d40

Revert D15459166: [pytorch][PR] add batch of string ops

ezyang committed 5 years ago

a21cf765

Remove duplicated _optimize_trace and use core (#20394)

lanpa committed 5 years ago

5952ca8d

move batchnorm and layernorm fusion to decompose (#20337)

wanchaol committed 5 years ago

871c9dcb

Quantized Conv2d operator (#20772)

dskhudia committed 5 years ago

cde611a6

add batch of string ops (#20826)

Elias Ellison committed 5 years ago

aebcd80a

make wildcards alias only each other (#20670)

suo committed 5 years ago

7aa3887f

Mark values entering containers as wildcards (#20556)

suo committed 5 years ago

90910fc6

Fix bug in exporting node with multiple outputs by scripting

BowenBao committed 5 years ago

28be521e

fix pow bug on overloads and clean up (#20824)

wanchaol committed 5 years ago

c2e3e79a

Allow both Variables and Tensors in c10 kernel interface (#20816)

smessmer committed 5 years ago

98928f4d

Add as_quantized_tensor (#20740)

jerryzh168 committed 5 years ago

9ea009fe

Change comparison ops result dtype to bool [Part1] (#20767)

izdeby committed 5 years ago

12bc81ae

Newer Older