strisunshinewentingwang/pytorch

Pull Requests Commits

Fix zero_grad on view grads

ssnl committed 4 years ago

Verified b45a569d

Fuse module enhancements (#26457)

raghuramank100 committed 4 years ago

dddae3f8

Support layout() in script

James Reed committed 4 years ago

6a4ca9ab

Add C++ torch::nn::HingeEmbeddingLoss (#27101)

jon-tow committed 4 years ago

209dc4c4

Adds Device Generic Precision Tests to test_torch.py (#26762)

Mike Ruberry committed 4 years ago

ea414e49

Updating submodules

svcscm committed 4 years ago

1d4d6b6f

Add control for observers in Fake-quantize module (#27113)

raghuramank100 committed 4 years ago

9e3ba355

Support for add relu functional module (#26612)

raghuramank100 committed 4 years ago

bdcaf633

Cuts test_torch.py runtime in half by marking four tests as slow (#26789)

Mike Ruberry committed 4 years ago

ec7913af

make class types callable (#26743)

Elias Ellison committed 4 years ago

990f4ca7

Set number of threads for operator_benchmarks (#27010)

dskhudia committed 4 years ago

bf7ebc5a

Fix `toIValue` dict iteration (#26856)

davidriazati committed 4 years ago

5e7549c3

separate out rpc to sync and async apis (#26570)

rohan-varma committed 4 years ago

fb8fd6dc

Module method destroy

IvanKobzarev committed 4 years ago

3e20d9c0

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

Older