strisunshinewentingwang/pytorch

Pull Requests Commits

run cpp tests for non-cuda builds in test_jit.py (#18826)

suo committed 5 years ago

0a4117a3

Fix the linter (#18842)

houseroad committed 5 years ago

100f95a3

Enforce single parent for script submodules (#18379)

zdevito committed 5 years ago

7e59c604

Allow ints, floats, and tensors in conditional (#18755)

Elias Ellison committed 5 years ago

b80a4fa2

Fix layernorm ad formula on weight and bias (#18233)

wanchaol committed 5 years ago

843e6234

Unify namespace of script::Module (#18378)

zdevito committed 5 years ago

0512e4e3

Step 1: Secretly add return_counts to unique, and refactor unique_dim for performance (#18648)

VitalyFedyunin committed 5 years ago

773ce4fb

Support replicating multi-GPU modules (#18687)

mrshenli committed 5 years ago

7ae0263e

wanchaol committed 5 years ago

eabd9eac

Remove `device_guard: False` from native_functions that don't have a … (#18803)

gchanan committed 5 years ago

862aff64

Switch our Linux machine AMI to a newer image. (#18433)

ezyang committed 5 years ago

cb959aa7

QTensor (#18230)

jerryzh168 committed 5 years ago

dfcd7b01

Enforce import order to make protobuf cpp implementation in python work (#18560)

Dmytro Dzhulgakov committed 5 years ago

3af2d6d9

Pin onnx ir_version to 4 (#18768)

houseroad committed 5 years ago

3b71f2e1

fix nccl compilation to make sure it compiles for architectures that pytorch compiles for (#18739)

soumith committed 5 years ago

8711df89

push magma init into lazyInitCUDA (#18527)

soumith committed 5 years ago

b5d8844b

For some files that are touched by the QTensor diff (#18765)

jerryzh168 committed 5 years ago

ed9724f3

Fix contiguous AD and Autogradzero inconsistency (#18633)

wanchaol committed 5 years ago

a21e256e

Added indexing for bool tensors and bool Indices (#18583)

izdeby committed 5 years ago

5950c1e8

add an assertion to check the param num (#18145)

houseroad committed 5 years ago

65dfe120

add Android NDK param to CI docker build script (#18782)

ljk53 committed 5 years ago

4afc067f

Upgrade mkldnn-bridge for dnnlowp support (#16308)

gujinghui committed 5 years ago

a7b82a44

add 'abs' builtin

TheCodez committed 5 years ago

46a68c1b

Fix dense Embedding to work with double backward (#9078)

kshitij12345 committed 5 years ago

0916b541

Highlight NCCL all_reduce and all_gather requirements (#18741)

mrshenli committed 5 years ago

c0ad6747

Updating submodules

svcscm committed 5 years ago

2658190d

Make it possible for users for select /Zi or /ZI over /Z7 when using MSVC (#18790)

peterjc123 committed 5 years ago

5e33085f

use optimization in D14020675 (#16945)

jspark1105 committed 5 years ago

06b7fe59

Add device and dtype to storage. (#18749)

gchanan committed 5 years ago

2113ea6f

Use non-legacy constructors for tensor deserialization. (#18750)

gchanan committed 5 years ago

a3da3653

Newer Older