strisunshinewentingwang/pytorch

Pull Requests Commits

[nn] Add allow_duplicate option for named_modules (#54812)

jerryzh168 committed 3 years ago

0a541e23

Implicit conversion from null tensor to NoneType (#55823)

tugsbayasgalan committed 3 years ago

b405e2ce

Set ThreadLocalState correctly in the autograd engine (#56174)

zou3519 committed 3 years ago

d2d11125

[package] fix error handling with allow_empty (#56190)

suo committed 3 years ago

8f683964

[optim] take kw-only argument for functional optim APIs (#56185)

wanchaol committed 3 years ago

46113876

[PyTorch Edge] Move torch::jit::mobile::_export_operator_list() from serialization/export_module.cpp to mobile/import.cpp (#56044)

dhruvbird committed 3 years ago

bd3c63ae

[iOS GPU] Use setTexture() rather than copyTexture() (#56069)

xta0 committed 3 years ago

94ce10f7

[DDP] Fixes flaky tests caused by incorrect floating-point comparison (#56192)

cbalioglu committed 3 years ago

42f5d660

.github: Add initial linux CI workflow (#55176)

samestep committed 3 years ago

7d410bc3

[PARAM] Param comms debug info (#55976)

chaekit committed 3 years ago

40039800

[tensorexpr] Add missing python bindings for NNC Stmts (#55570)

huiguoo committed 3 years ago

bde53cfd

ns for fx: add test for op relationship coverage (#55837)

vkuzo committed 3 years ago

f59244ec

ns for fx: move pattern utils to separate file (#55805)

vkuzo committed 3 years ago

c8209a73

ns for fx: make get_reversed_fusions reuse quantization fusions (#55803)

vkuzo committed 3 years ago

b4611045

ns for fx: add qat tests cases for shadowed activations (#55614)

vkuzo committed 3 years ago

84b5f67d

ns for fx: qat test cases for unshadowed activations (#55508)

vkuzo committed 3 years ago

37fbc069

ns for fx: extend functional weight extraction testing to QAT (#55507)

vkuzo committed 3 years ago

f6a3936a

ns for fx: add qat handling for weight extraction (#55506)

vkuzo committed 3 years ago

1cbc4023

ns for fx: make NSTracer inherit from QuantizationTracer (#55505)

vkuzo committed 3 years ago

3786c271

ns for fx: change node I/O determination to strict allowlist (#55434)

vkuzo committed 3 years ago

5ad3bc71

[kineto] deprecate metdata args from ClientTraceActivity (#55988)

chaekit committed 3 years ago

1ca51f0f

Python API for Vitals (#53238)

Victor Bittorf committed 3 years ago

52f1a07b

Port all unary float functions to structured (#56082)

ezyang committed 3 years ago

f17c9ea2

Change all unary functions stubs to use TensorIteratorBase& (#56078)

ezyang committed 3 years ago

cfc97162

remove annoying warnings from common_nn.py (#55982)

Chillee committed 3 years ago

3c4e1cd1

Add cost inference for MulGradient operator

Chao Kong committed 3 years ago

ff1498e6

port addmv to structured kernels (#55746)

ngimel committed 3 years ago

3fbca31b

Reland: D27652485: [nnc] Enable CPU fusion only when num_threads == 1" (#56120)

bertmaher committed 3 years ago

8e82e932

[reland][ROCm] use hiprtc precompiled header (#55965)

jeffdaily committed 3 years ago

e1752ffa

Fix ChanelShuffle named tensor warnings (#55911)

h6197627 committed 3 years ago

f02454f9

Newer Older