strisunshinewentingwang/pytorch

Pull Requests Commits

Move THTensor_compute_stride to ATen (#23045)

jerryzh168 committed 5 years ago

a03205ed

Add fused modules in nn._intrinsic (#23085)

jerryzh168 committed 5 years ago

0d8324b1

Quantized concatenation (+fused relu). (#21749)

Zafar Takhirov committed 5 years ago

47af41fe

Moving np function to test area

z-a-f committed 5 years ago

9f4df63c

Conv module (#23084)

jerryzh168 committed 5 years ago

77353636

Fbgemm fp16 tensor support (#23101)

Yinghai Lu committed 5 years ago

b964bdb5

Fix workspace destruction ordering (#23096)

Yinghai Lu committed 5 years ago

2a8d5a13

Include module names in recursive error stacks (#22921)

davidriazati committed 5 years ago

79c4f83f

Quantization aware training in eager mode (#23082)

jerryzh168 committed 5 years ago

7cc029cb

Add initial support for serializing classes

zdevito committed 5 years ago

c09e9225

BlackBoxPredictor OSS: open-source NQL and custom transforms (#22877)

salexspb committed 5 years ago

6334edc2

fix overspecializing constants in compilation (#22816)

Elias Ellison committed 5 years ago

f2f3e8ad

Adding more binary documentation

pjh5 committed 5 years ago

a302821c

jjsjann123 committed 5 years ago

a28ffaf3

Only import PIL when needed (#23023)

orionr committed 5 years ago

818828e8

Pass variable_list of inputs to _wrap_outputs

mal committed 5 years ago

44493a62

add break continue to docs

Elias Ellison committed 5 years ago

2ee0f0bc

Remove deprecated linear algebra functions (and methods) (#22841)

vishwakftw committed 5 years ago

6dfecc7e

Delete aten/src/ATen/out.txt (#23050)

xuhdev committed 5 years ago

61a683c2

Fix get_all_math_dtypes for device='cuda' retuning None (#23028)

Morgan Funtowicz committed 5 years ago

5417ddbd

Revert D16199356: [qat] Quantization aware training in eager mode

soumith committed 5 years ago

84c2c89e

Revert D16274792: [qat] Conv module

soumith committed 5 years ago

f19aa12a

Revert D16349133: [quant] Add fused modules in nn._intrinsic

soumith committed 5 years ago

c362e72d

Revert D16373996: [fix] conv module missing return

soumith committed 5 years ago

2401a05a

BERT CPU performance optimization: use mkldnn for nn.Linear() when input is dense layout (#21851)

mingfeima committed 5 years ago

25f0dc34

fix error message

orena1 committed 5 years ago

12ac9171

conv module missing return (#23058)

jerryzh168 committed 5 years ago

cdfdeb74

Use ProfiledTensorType in peephole.cpp

Krovatkin committed 5 years ago

66019780

Updating submodules

svcscm committed 5 years ago

d153b0b5

Fix TBB build for older versions of cmake

Ilia Cherniavskii committed 5 years ago

23badc60

Newer Older