strisunshinewentingwang/pytorch

Pull Requests Commits

Script for full android build to aars; script to run android tests (#26833)

IvanKobzarev committed 4 years ago

47cd15c6

Add temporary torch::k{name} enum declarations (#27051)

Will Feng committed 4 years ago

27d4b34e

Make static dispatch turn off variable before entering the kernel. (#26908)

ezyang committed 4 years ago

9159a601

Updating submodules

svcscm committed 4 years ago

e75a4934

Return NotImplemented from tensor arithmetic operators (#26507)

peterbell10 committed 4 years ago

585a5975

Don't apply should_run to the nightly/postnightly branches. (#27061)

ezyang committed 4 years ago

4a8d899e

Work around a gcc-7 bug in building Debug version of Sleef (#26993)

xuhdev committed 4 years ago

73347e0a

Revert D17666050: [pytorch][PR] Fixed seek offset size to 64bit.

ezyang committed 4 years ago

b16358b2

Add tuple constructor + to<std::tuple<Args...>> (#26668)

Bram Wasti committed 4 years ago

e0833873

Integrate prepacked workaround in QuantFusion (#26939)

jerryzh168 committed 4 years ago

e367f605

serialize autograd ops into its own namespace (#26761)

wanchaol committed 4 years ago

a252aee8

Fold prepacked weight into module (#26579)

jerryzh168 committed 4 years ago

d91e490a

Back out "Add int8 resize nearest 3d op in DNNLOWP" (#27056)

philippv committed 4 years ago

3d41ad50

Fixed seek offset size to 64bit. (#27047)

yn4bit committed 4 years ago

1afe3fc0

peterjc123 committed 4 years ago

1eaa9f89

Make nonzero non differentiable as it supposed to be (#26980)

VitalyFedyunin committed 4 years ago

275e0c1c

Default observer and fake-quant for backends (#26627)

raghuramank100 committed 4 years ago

d5298b6e

Emulate weight and activation only quant with fake quant, numerics test (#26625)

raghuramank100 committed 4 years ago

32b0e8c9

Quantization aware training: Freeze batch norm support (#26624)

raghuramank100 committed 4 years ago

84ee8ace

Per channel fake quant (#26623)

raghuramank100 committed 4 years ago

7dc70757

Add boxed fallback (#26368)

ezyang committed 4 years ago

7c465aae

make cudnn rnn respect current stream (#27026)

ngimel committed 4 years ago

2bcbc150

Integrate FC fp16 exporter into Dper2 (#26582)

Xing Wang committed 4 years ago

a1513dce

C++ API parity: AdaptiveAvgPool3d

pbelevich committed 4 years ago

1a3997e0

Improve repr for quantized modules

James Reed committed 4 years ago

4d7bec5f

Migrate le/gt/ge/eq/ne from the TH to Aten. Added support of type promotion. (#27017)

ifedan committed 4 years ago

ee2c79d6

Turn Caffe2 CUDA 9.1 + py2 to CUDA 10.1 + py3 (#26835)

houseroad committed 4 years ago

9e1cf99a

Per-channel baseline (#26516)

raghuramank100 committed 4 years ago

2ccbdb79

C++ API parity: AdaptiveAvgPool2d

pbelevich committed 4 years ago

a31fd5ea

C++ API parity: AdaptiveAvgPool1d

pbelevich committed 4 years ago

7d58060f

Newer Older