strisunshinewentingwang/pytorch

Pull Requests Commits

Add DDP+RPC tutorial to RPC docs page. (#42828)

pritamdamania committed 4 years ago

a6b69fdd

make deadline=None for all numerics tests (#43014)

Hector Yuen committed 4 years ago

3544f60f

Fix to Learnable Fake Quantization Op Benchmarking (#43018)

Paul Shao committed 4 years ago

8b5642a7

Enable torch.utils typechecks (#42960)

malfet committed 4 years ago

6753157c

Add executor and fuser options to the fastrnn test fixture (#42946)

bertmaher committed 4 years ago

eb47940c

Update ort-nightly version to dev202008122 (#43019)

malfet committed 4 years ago

fd5ed4b6

[quant] Make PerChannel Observer work with float qparams (#42690)

supriyar committed 4 years ago

816d37b1

[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

Newer Older