strisunshinewentingwang/pytorch

Pull Requests Commits

Back out "Back out "[pytorch][PR] Move thnvrtc and DynamicLibrary to ATen"" (#22794)

ezyang committed 5 years ago

535c5540

Remove tensor_data() call in Python Variable() and nn.Parameter() constructors (#22821)

Will Feng committed 5 years ago

317cf7c8

Make the signature of fill_out consistent with fill_.

xuhdev committed 5 years ago

14e8fb70

Move the body of fill_kernel_impl into fill_kernel_cuda

xuhdev committed 5 years ago

1c266c27

Move fill and fill_diagonal to Fill.cpp, Fill.h, and FillKernel.{cpp,cu}

xuhdev committed 5 years ago

fc297b8e

make_variable consumes the Tensor if it only has one reference

James Reed committed 5 years ago

815e73bc

Temporarily skip mypy-0.720 to unbreak master type checks

bddppq committed 5 years ago

b5fa9a34

lly-zero-one committed 5 years ago

1a93b968

Revert D16238204: Revert D16224780: [pytorch][PR] [ROCm] MIOpen integration into pytorch RNN operators

bddppq committed 5 years ago

92468f0a

Revert D16037021: Support GRU module quantization in Pytorch

Karl Ostmo committed 5 years ago

da315a4e

Revert D16224780: [pytorch][PR] [ROCm] MIOpen integration into pytorch RNN operators

Karl Ostmo committed 5 years ago

fcfefc34

Transfer Learning: Caffe2 load op changes to return shape inference (#22829)

pradeepdfb committed 5 years ago

ead11932

Support GRU module quantization in Pytorch

lly-zero-one committed 5 years ago

d8c1b861

Get rid of torch.mean shape analysis

James Reed committed 5 years ago

ba9d559a

Use Linear Operator with fp16 weights in JIT (#22323)

mingzhe09088 committed 5 years ago

9eb03933

Support Linear operation with fp16 weights in ATen (#22023)

mingzhe09088 committed 5 years ago

573d9e69

Revert D16204820: [pytorch][PR] optimize topk on cpu using parallel and partial sort

Karl Ostmo committed 5 years ago

35ee4bf4

add support for breaks and continues (#21692)

Elias Ellison committed 5 years ago

cf2889ad

PyTorch export to ONNX Opset 7 and 8 - Cont (#22421)

BowenBao committed 5 years ago

b3147bc6

MIOpen integration into pytorch RNN operators (#22774)

ashish committed 5 years ago

9f8e2c06

Speeds up fast-path for 1D tensors (#22756)

colesbury committed 5 years ago

30e03df6

avoid kernel launches for zero-sized tensor inputs

wanchaol committed 5 years ago

02bc06a6

Make PythonArgs::tensor and PythonArgs::scalar faster (#22782)

colesbury committed 5 years ago

b1b65f34

optimize topk on cpu using parallel and partial sort (#19736)

mingfeima committed 5 years ago

10c14ad1

Speed up TensorIterator::compute_strides a little (#22779)

colesbury committed 5 years ago

fc23d7f3

Initiate checkCuda90Bug warning (#22757)

Guanheng Zhang committed 5 years ago

f266a63e

Revert D16222539: [pytorch][PR] Let users pass CMake-specific options starting with CMAKE_ to CMake.

ezyang committed 5 years ago

ccb28939

Let users pass CMake-specific options starting with CMAKE_ to CMake. (#22776)

xuhdev committed 5 years ago

612eed31

add new tests to benchmark_all_test (#22787)

mingzhe09088 committed 5 years ago

7eb03193

make custom op work in OSS environment (#22781)

mingzhe09088 committed 5 years ago

1878800f

Newer Older