strisunshinewentingwang/pytorch

Pull Requests Commits

dedupe test skipping in common_distributed and test_distributed (#38078)

rohan-varma committed 4 years ago

45010833

Use py::pickle in RRef pickling pybind code (#38147)

rohan-varma committed 4 years ago

e109ff63

Bind VariableFunctions as a module, not a class with static methods. (#38136)

ezyang committed 4 years ago

30f4064c

Add minimal skeleton for _C type stubs, delete torch.autograd stub (#38080)

ezyang committed 4 years ago

7e9af67c

[TensorExpr] Add print functions for Tensor and Function. (#38175)

Mikhail Zolotukhin committed 4 years ago

464e5a6c

Automatic update of fbcode/onnx to 79a7e0df7e86e0f32e7a05f563b24a566540c18b (#38106)

houseroad committed 4 years ago

81817116

Consolidate builtin/python_udf RPC to return ivalue::Future like torchscript RPC does (#35154)

xush6528 committed 4 years ago

3d027986

Remove uses of type() part 2 (#38140)

Ailing Zhang committed 4 years ago

86d28706

Unboxing uses if_constexpr instead of SFINAE (#38145)

smessmer committed 4 years ago

16e62f93

[TorchScript] Explicitly disallow del with more than 1 operand. (#38089)

ycao committed 4 years ago

ae534dc9

[quant] Disable qnnpack test when TSAN is enabled (#38153)

supriyar committed 4 years ago

13847638

Fix overflow in torch.remainder when dividend is very large (#37758)

xwang233 committed 4 years ago

63b1ae69

s/callUnboxed/call/ (#37999)

Basil Hosmer committed 4 years ago

fdc40616

Add test jobs on CPU agents for CUDA builds on Windows (#37904)

peterjc123 committed 4 years ago

55de7c3b

[JIT]Support List variable in adv indexing. (#37966)

Ailing Zhang committed 4 years ago

e84aa021

Vectorize non-persistent Softmax kernels (#36485)

slayton58 committed 4 years ago

c879c6fb

Migrate OwnerRRef value store to generic torch Future (#38143)

xush6528 committed 4 years ago

615235fc

Add documentation for FeatureAlphaDropout (#36295)

Donna Choi committed 4 years ago

ca2206d0

Fix a minor typo in DistanceOpsKernel.cpp (#37596)

yoshnary committed 4 years ago

c13dc2ca

[TensorExpr] Fix a bug in the IR Simplifier that could introduce a division by zero (#38055)

nickgg committed 4 years ago

ad433e20

int8 fc with tests (#38017)

Hector Yuen committed 4 years ago

9957db22

Add documentation for nn.Hardsigmoid and nn.functional.hardsigmoid. (#38120)

zou3519 committed 4 years ago

172bcdb8

Dont store redundant packed params in dynamic quantized RNN (#38134)

James Reed committed 4 years ago

41572116

[TensorExpr] Don't include aten::rand_like to TE fusion groups since we can't handle rand+broadcast case yet. (#38132)

Mikhail Zolotukhin committed 4 years ago

4784af1d

fix compilation error with gcc 5.5 (#38112)

cloudhan committed 4 years ago

6e1e2a60

`unfold_backward` gets its own kernel (#36612)

nikitaved committed 4 years ago

a7c29dbf

[quant][graphmode][refactor] Split quantization.cpp (#37975)

jerryzh168 committed 4 years ago

0ed7fc58

[quant][graphmode][refactor] Remove unused code in quantization.cpp (#37974)

jerryzh168 committed 4 years ago

ff9a809c

Back out "Revert D20229168: [quantization] Use torchbind for Linear PackedParams" (#38101)

James Reed committed 4 years ago

c1e7758b

[TensorPipe] Do not require user to provide worker name-to-rank map (#38052)

lw committed 4 years ago

91f451a5

Newer Older