strisunshinewentingwang/pytorch

Pull Requests Commits

ONNX Export All Cases of Softmax

lara-hdr committed 5 years ago

1ec1db47

Added bool and half support for resize_as_ and view methods (#18821)

izdeby committed 5 years ago

b4d2df1f

Automatic update of fbcode/onnx to 079c2639f9bb79b1774d1e3bfa05b0c093816ca7 (#18841)

houseroad committed 5 years ago

bb16e8da

Actually model scalar type promotion in shape analysis (#18811)

James Reed committed 5 years ago

33f4751f

Add a .ctags.d/ toplevel directory (#18827)

mxw committed 5 years ago

d108a1ab

Wanwannodao committed 5 years ago

8ca9ba17

Add support for group ConvTranspose (#18794)

xiaomengy committed 5 years ago

b145dcca

Disallow changing the device of a tensor via set_. (#18832)

gchanan committed 5 years ago

8732a1b4

U/kostmo/win test offload scripts

kostmo committed 5 years ago

15b318de

Revert D14603722: Enforce single parent for script submodules

zdevito committed 5 years ago

f97eb8d9

Fix deviceCount on FakeGuardImpl. (#18745)

ezyang committed 5 years ago

52a3a514

Stop swapping in Storages of the wrong device for Tensors. (#18831)

gchanan committed 5 years ago

486fae56

Pass ScalarType separately from Type in python constructors

Roy Li committed 5 years ago

d70c6f23

Store ScalarType and Backend instead of Type in TensorIterator

Roy Li committed 5 years ago

f5741eb8

Introduce DeprecatedTypeProperties class (#17991)

Roy Li committed 5 years ago

c705d9eb

Fix to handle null strides in DLPack tensor (#18510)

bwasti committed 5 years ago

095f88e0

Add shape inference function for Split (#18838)

Yinghai Lu committed 5 years ago

e5e2110a

Fix the duplication problem in _unique_state_dict (#18139)

houseroad committed 5 years ago

0c237f13

fold col offset into bias; optimize A symmetric quant (#17026)

jspark1105 committed 5 years ago

fa0ad057

fix flake8 lint (#18835)

suo committed 5 years ago

72913a55

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

Newer Older