Question about

[gaohongfein]

I have succeed compile the structural_losses. But it can still not import the Chamfer-EMD. Do you know why?
External Losses (Chamfer-EMD) were not loaded.
External Losses (Chamfer-EMD) cannot be loaded. Please install them first.

make: 放弃循环依赖 tf_approxmatch_g.cu <- tf_approxmatch_g.cu.o 。
/usr/local/cuda-8.0/bin/nvcc -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++11 -c -o tf_approxmatch_g.cu.o tf_approxmatch_g.cu -I /home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include -DGOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -O2
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
g++ -std=c++11 tf_approxmatch.cpp tf_approxmatch_g.cu.o -o tf_approxmatch_so.so -shared -fPIC -I /home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include -lcudart -L /usr/local/cuda-8.0/lib64 -O2 -D_GLIBCXX_USE_CXX11_ABI=0
make: 放弃循环依赖 tf_nndistance_g.cu <- tf_nndistance_g.cu.o 。
/usr/local/cuda-8.0/bin/nvcc -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++11 -c -o tf_nndistance_g.cu.o tf_nndistance_g.cu -I /home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include -DGOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -O2
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1254): warning: calling a constexpr host function("real") from a host device function("abs") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1254): warning: calling a constexpr host function("imag") from a host device function("abs") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1254): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1254): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1259): warning: calling a constexpr host function("real") from a host device function("abs") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1259): warning: calling a constexpr host function("imag") from a host device function("abs") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1259): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/../../../Eigen/src/Core/MathFunctions.h(1259): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/src/Tensor/TensorRandom.h(133): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/src/Tensor/TensorRandom.h(138): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/src/Tensor/TensorRandom.h(212): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

/home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include/unsupported/Eigen/CXX11/src/Tensor/TensorRandom.h(217): warning: calling a constexpr host function from a host device function is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this.

g++ -std=c++11 tf_nndistance.cpp tf_nndistance_g.cu.o -o tf_nndistance_so.so -shared -fPIC -I /home/gao/tf2/local/lib/python2.7/site-packages/tensorflow/include -lcudart -L /usr/local/cuda-8.0/lib64 -O2 -D_GLIBCXX_USE_CXX11_ABI=0

[Vahe1994]

What cuda and tensorflow version do you using?
Try to import nn_distance directly,I think you will see something like this "undefined symbol: _ZTIN10tensorflow8OpKernelE" in jupyter notebook. If you see that error than make a look here #3 .I hope it will help you.

[yuqli]

I had problems importing nn_distance too. But then I changed tensorflow to version 1.3, as well as nvcc version from 10.0 to 8.0, then everything seems to work fine. I think it's some problem in the make process that eventually leads to the import failure.

[triple-tam]

Has anyone used

• CUDA 9.0
• cuDNN v7
• tensorflow 1.5.0
  and had success compiling the loss operators?