strisunshinewentingwang/pytorch

Pull Requests Commits

[quant] Create PerRowQuantizer for floating point scale and zero_point (#42612)

supriyar committed 4 years ago

6f844684

collect_env.py: Print CPU architecture after Linux OS name (#42961)

malfet committed 4 years ago

0ff51acc

Do not ignore `torch/__init__.pyi` (#42958)

malfet committed 4 years ago

ebc7ebc7

[NNC] Fix some bugs in Round+Mod simplification (#42934)

nickgg committed 4 years ago

6fb5ce55

update clone doc (#42931)

albanD committed 4 years ago

f03f9ad6

[tensorexpr] Autograd for testing (#42548)

bwasti committed 4 years ago

ba9025bc

Revert D22856816: [quant][fix] Remove activation_post_process in qat modules

zou3519 committed 4 years ago

607e49cc

Enroll TensorPipe agent in C++-only E2E test (#42680)

lw committed 4 years ago

8493b0d5

Remove Python dependency from TensorPipe RPC agent (#42678)

lw committed 4 years ago

c88d3a5e

[Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

generatedunixname89002005287564 committed 4 years ago

d39cb84f

[quant][pyper] Make offsets an optional paramter in the qembedding_bag op (#42924)

supriyar committed 4 years ago

c9dcc833

[quant][fix] Remove activation_post_process in qat modules (#42343)

jerryzh168 committed 4 years ago

8cb42fce

Remove impl_unboxedOnlyKernel (#42841)

smessmer committed 4 years ago

7a7424bf

Allow Tensor& in the unboxing logic (#42712)

smessmer committed 4 years ago

20e0e54d

Add missing type annotation for Tensor.ndim (#42909)

KKawamura1 committed 4 years ago

5d2e9b6e

Add a microbenchmark for LSTM elementwise portion (#42901)

bertmaher committed 4 years ago

b8ae563c

Fix TE microbenchmark harness to use appropriate fuser/executor (#42900)

bertmaher committed 4 years ago

33d209b5

Speed up CUDA kernel launch when block/thread extents are statically known (#42899)

bertmaher committed 4 years ago

1adeed27

Revert D22994446: [pytorch][PR] CUDA reduction: allow outputs to have different strides

ngimel committed 4 years ago

f373cda0

Update cuda init docstring to improve clarity (#42923)

davanstrien committed 4 years ago

86841f5f

[FX] Add interface to reject nodes (#42865)

James Reed committed 4 years ago

0134deda

Implement hypot (#42291)

muthuArivoli committed 4 years ago

92885ebe

Implemented non-named version of unflatten (#42563)

heitorschueroff committed 4 years ago

62bd2dde

CUDA reduction: allow outputs to have different strides (#42649)

zasdfgbnm committed 4 years ago

7f3f5020

[jit] Scaffold a static runtime (#42753)

bwasti committed 4 years ago

ada8404f

[pytorch] BUCK build for Vulkan backend

IvanKobzarev committed 4 years ago

59f86923

Use `string(APPEND FOO " bar")` instead of `set(FOO "${FOO} bar") (#42844)

malfet committed 4 years ago

ea65a568

Revert D22898051: [pytorch][PR] Fix freeze_module pass for sharedtype

zou3519 committed 4 years ago

3d3752d7

Improve calling backward() and grad() inside vmap error messages (#42876)

zou3519 committed 4 years ago

bda00076

Fix get_writable_path (#42895)

malfet committed 4 years ago

5c39146c

Newer Older