| .. |
|
neon
|
Use scalar implementation instead of neon implementation to avoid out of range memory access in the tail conv3x3.
|
8 years ago |
|
nnpack
|
Fix some bugs.
|
8 years ago |
|
BlockExpandOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
BlockExpandOpTest.cpp
|
Simplify BlockExpandOpTest.
|
8 years ago |
|
BufferArg.cpp
|
Further address Daoyuan's comments, clean the code.
|
8 years ago |
|
BufferArg.h
|
merge
|
8 years ago |
|
BufferArgTest.cpp
|
Fix comparison between signed and unsigned values
|
8 years ago |
|
CMakeLists.txt
|
MulValue --> ScaleSubRegion
|
8 years ago |
|
ContextProjectionOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
ContextProjectionOp.h
|
rewrite unit test using new Function Test.
|
8 years ago |
|
ContextProjectionOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
ContextProjectionOpTest.cpp
|
Modify FunctionCompare to Compare2Function to support comparison of two CPU functions.
|
8 years ago |
|
ConvOp.h
|
support exconv dilation
|
8 years ago |
|
ConvOpTest.h
|
regenerate the proto for support the dilation
|
7 years ago |
|
CosSimOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
CosSimOp.h
|
merge Daoyuan's FuncArg, address one of the comments.
|
8 years ago |
|
CosSimOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
CosSimOpTest.cpp
|
Modify FunctionCompare to Compare2Function to support comparison of two CPU functions.
|
8 years ago |
|
CropOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
CropOp.h
|
1. Reading image shape from input data instead of image_config
|
8 years ago |
|
CropOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
CropOpTest.cpp
|
fix unittest of crop layer
|
8 years ago |
|
CrossMapNormalOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
CrossMapNormalOp.h
|
follow some comments with issue 973
|
8 years ago |
|
CrossMapNormalOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
CrossMapNormalOpTest.cpp
|
Simplify CrossMapNormalOpTest.
|
8 years ago |
|
DepthwiseConvOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
DepthwiseConvOp.h
|
support inputchannels != outputchannels of depthwiseconv
|
8 years ago |
|
DepthwiseConvOpGpu.cu
|
Remove the header files that do not need to be included.
|
8 years ago |
|
DepthwiseConvOpTest.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
EigenGemm.cpp
|
Enable the case N != ldc in EigenBlasGemm. (#5976)
|
7 years ago |
|
Function.cpp
|
Using linb::any/std::any instead of FunctionConfig
|
8 years ago |
|
Function.h
|
fix PadOp bug on Gpu
|
8 years ago |
|
FunctionTest.cpp
|
Fix more warnings about comparison between signed and unsigned values
|
8 years ago |
|
FunctionTest.h
|
CallBack --> Callback
|
8 years ago |
|
GemmConvOp.cpp
|
Follow comments.
|
7 years ago |
|
GemmConvOpTest.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
GemmFunctor.cpp
|
Change class to struct in GemmFunctor to avoid errors on special compilers
|
8 years ago |
|
GemmFunctor.h
|
Add EigenGemm.
|
8 years ago |
|
GruFunctor.h
|
Change the definition of vmlaq_laneq_f32 from template function to macro.
|
8 years ago |
|
Im2Col.h
|
Follow comments.
|
7 years ago |
|
Im2ColOp.cpp
|
im2col cpu gpu dilation support
|
8 years ago |
|
Im2ColOpGpu.cu
|
im2col cpu gpu dilation support
|
8 years ago |
|
Im2ColTest.cpp
|
Follow comments.
|
7 years ago |
|
MulOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
MulOp.h
|
Daoyuan's comments.
|
8 years ago |
|
MulOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
MulOpTest.cpp
|
Modify FunctionCompare to Compare2Function to support comparison of two CPU functions.
|
8 years ago |
|
NaiveConvOp.cpp
|
FIX: cppint code style
|
8 years ago |
|
PadOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
PadOp.h
|
Add Error in FuncConfig.
|
8 years ago |
|
PadOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
PadOpTest.cpp
|
Modify FunctionCompare to Compare2Function to support comparison of two CPU functions.
|
8 years ago |
|
RowConvOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
RowConvOp.h
|
Fix bug and Python API.
|
8 years ago |
|
RowConvOpGpu.cu
|
ClangFormat for proto and cuda
|
8 years ago |
|
RowConvOpTest.cpp
|
Fix RowConvOpTest use CpuGpuFuncCompare.
|
8 years ago |
|
ScaleSubRegionOp.cpp
|
MulValue --> ScaleSubRegion
|
8 years ago |
|
ScaleSubRegionOp.h
|
MulValue --> ScaleSubRegion
|
8 years ago |
|
ScaleSubRegionOpGpu.cu
|
MulValue --> ScaleSubRegion
|
8 years ago |
|
ScaleSubRegionOpTest.cpp
|
Remove MulValu* and reduce time cost for unit test.
|
8 years ago |
|
SwitchOp.cpp
|
Use PADDLE_WITH_CUDA instead of PADDLE_WITH_GPU
|
8 years ago |
|
SwitchOp.h
|
1. delete PixelSoftmaxLayer and add SwitchOrderLayer
|
8 years ago |
|
SwitchOpGpu.cu
|
Fix format
|
8 years ago |
|
SwitchOpTest.cpp
|
Add pixel softmax layer for FCN model
|
8 years ago |
|
TensorShape.h
|
update code
|
8 years ago |
|
TensorShapeTest.cpp
|
Fix some build warnings
|
8 years ago |
|
TensorType.h
|
Further address Daoyuan's comments, clean the code.
|
8 years ago |
|
TensorTypeTest.cpp
|
Fix comparison between signed and unsigned values
|
8 years ago |
hedaoyuan