strisunshinewentingwang/pytorch

Pull Requests Commits

Add ROCm-specific half_support_literal for JIT. (#38899)

jeffdaily committed 4 years ago

ccab1421

Updates torchvision version (#38848)

mruberry committed 4 years ago

0d649efb

Fix histc with empty tensor error (#38987)

ShawnZhong committed 4 years ago

12c219de

[c2] cuda impl for WeightScale op (#38712)

Wakeupbuddy committed 4 years ago

c40a79a0

Revert D21681838: add eq.str op to lite interpreter

rohan-varma committed 4 years ago

224ce03e

Fix max_pool2d nchw backward bug (#38953)

xwang233 committed 4 years ago

e4a3c584

Port TH cum{sum,prod}_cuda to ATen (#36458)

kshitij12345 committed 4 years ago

0ff1aa90

[vulkan] Fix python overrides tests for is_vulkan_available (#39016)

IvanKobzarev committed 4 years ago

996b6a3d

Fix max_pool2d for returning wrong shape with return_indices=True on cuda (#38992)

xwang233 committed 4 years ago

583ff947

[vulkan] Fix Bazel build, add aten/native/vulkan/stub/*.cpp (#39018)

IvanKobzarev committed 4 years ago

c8237530

Remove reference of CUDA < 9.2 (#38977)

ShawnZhong committed 4 years ago

fc4dfbf7

move int8 fc operators and dependencies (#38935)

Hector Yuen committed 4 years ago

108321dc

Use a temporary file during ReducerTest (#39004)

diego-plan9 committed 4 years ago

4d1df74c

Disable some unsupported module for 32-bit build (#38950)

peterjc123 committed 4 years ago

1fef2075

Expose VC_YEAR in Windows binary test jobs (#38957)

peterjc123 committed 4 years ago

81daadf6

Updates assertEqual to require atol and rtol, removes positional atol (#38872)

Mike Ruberry committed 4 years ago

6ddca30b

add eq.str op to lite interpreter (#38859)

linbinyu committed 4 years ago

341fd63f

[Mobile GPU][Integration] Vulkan backend integration (#36491)

IvanKobzarev committed 4 years ago

b460465a

Use workspace to persist and restore images for Windows CI build and … (#38971)

guyang3532 committed 4 years ago

1fa0bb6d

Updating submodules

svcscm committed 4 years ago

f07a60fc

improve accuracy of logsoftmax computation on cuda (#38945)

ngimel committed 4 years ago

c34b3332

`torch.pow` Add type promotion support and fix issue with __rpow__ (#37098)

nairbv committed 4 years ago

389e16c3

Rename `torch._C.Generator` to `torch.Generator` (#38773)

ShawnZhong committed 4 years ago

ba3893e7

Update TensorPipe submodule (#38923)

lw committed 4 years ago

b8f2ecbf

Update ShipIt sync

yns88 committed 4 years ago

5749ef75

[PATCH] Migrate min from THC to ATen and remove _min (#38440)

zasdfgbnm committed 4 years ago

7e6f6f52

[pytorch] expose __ldg(const Half* ptr) to Clang in host mode (#38151)

Igor Sugak committed 4 years ago

d035d050

Use old circleci windows image for both CPU and CUDA (#38909)

peterjc123 committed 4 years ago

f3f3097a

[JIT] Factor out aliases to separate test (#38746)

Elias Ellison committed 4 years ago

cd5d7a34

[JIT] Normalize op aliases (#38735)

Elias Ellison committed 4 years ago

f90dc741

Newer Older