How to set makefile config?

[Sarah20187]

I can install successfully. But when I make bilateralNN, the cudnn parts always have errors. I wonder weather I need to set the makefile config? Errors：

/raid5/sarah/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_tanh_layer.cu(13): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNTanHLayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (43): here

/raid5/sarah/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_tanh_layer.cu(34): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNTanHLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (43): here

/raid5/sarah/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_tanh_layer.cu(13): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNTanHLayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (43): here

I searched on google that this problem is caused by cudnn version. However, I tried cudnn v4,5,6 by cmake -DCUDNN_INCLUDE=/home/sarah/dependencies/cudnn_v6_7.5/cuda/include/ -DCUDNN_LIBRARY=/home/sarah/dependencies/cudnn_v6_7.5/cuda/lib64 .. but this problem still remains.

figuration FYI

loading initial cache file /raid5/sarah/bilateralNN/build/tmp_caffe_clone/tmp/CaffeUpstream-cache-.cmake -- Boost version: 1.54.0 -- Found the following Boost libraries: -- system -- thread -- filesystem -- Found gflags (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libgflags.so) -- Found glog (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libglog.so) -- Found PROTOBUF Compiler: /usr/bin/protoc -- Found lmdb (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/liblmdb.so) -- Found LevelDB (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libleveldb.so) -- Found Snappy (include: /usr/include, library: /usr/lib/libsnappy.so) -- CUDA detected: 7.5 -- Found cuDNN: ver. 6.0.21 found (include: /home/sarah/dependencies/cudnn_v6_7.5/cuda/include, library: /home/sarah/dependencies/cudnn_v6_7.5/cuda/lib64) -- Added CUDA NVCC flags for: sm_52 -- OpenCV found (/home/sarah/anaconda2/share/OpenCV) -- Found Atlas (include: /usr/include, library: /usr/lib/libatlas.so) -- NumPy ver. 1.11.3 found (include: /home/sarah/anaconda2/lib/python2.7/site-packages/numpy/core/include) -- Boost version: 1.54.0 -- Found the following Boost libraries: -- python -- Could NOT find Doxygen (missing: DOXYGEN_EXECUTABLE) -- -- Caffe Configuration Summary -- General: -- Version : 1.0.0-rc3 -- Git : rc3-62-ga1c81ac-dirty -- System : Linux -- C++ compiler : /usr/bin/c++ -- Release CXX flags : -O3 -DNDEBUG -fPIC -Wall -Wno-sign-compare -Wno-uninitialized -- Debug CXX flags : -g -fPIC -Wall -Wno-sign-compare -Wno-uninitialized -- Build type : Release -- -- BUILD_SHARED_LIBS : ON -- BUILD_python : ON -- BUILD_matlab : OFF -- BUILD_docs : ON -- CPU_ONLY : OFF -- USE_OPENCV : ON -- USE_LEVELDB : ON -- USE_LMDB : ON -- ALLOW_LMDB_NOLOCK : OFF -- -- Dependencies: -- BLAS : Yes (Atlas) -- Boost : Yes (ver. 1.54) -- glog : Yes -- gflags : Yes -- protobuf : Yes (ver. 2.5.0) -- lmdb : Yes (ver. 0.9.10) -- LevelDB : Yes (ver. 1.15) -- Snappy : Yes (ver. 1.1.0) -- OpenCV : Yes (ver. 3.2.0) -- CUDA : Yes (ver. 7.5) -- -- NVIDIA CUDA: -- Target GPU(s) : Auto -- GPU arch(s) : sm_52 -- cuDNN : Yes (ver. 6.0.21) -- -- Python: -- Interpreter : /home/sarah/anaconda2/bin/python2.7 (ver. 2.7.13) -- Libraries : /usr/lib/x86_64-linux-gnu/libpython2.7.so (ver 2.7.13) -- NumPy : /home/sarah/anaconda2/lib/python2.7/site-packages/numpy/core/include (ver 1.11.3) -- -- Documentaion: -- Doxygen : No -- config_file :
-- -- Install: -- Install path : /raid5/sarah/bilateralNN/build/bin/install -- -- Configuring done -- Generating done -- Build files have been written to: /raid5/sarah/bilateralNN/build/bin

[raffienficiaud]

This seems to be a Caffee problem. Which version of Caffee are you using? Is this the one that has been advertised on the main readme? (a1c81aca641e5b16f3e2007be07dfdedc072606e)

[Sarah20187]

Yes，i use the first method，which download the caffe version you suggested itself. On Sat, 20 May 2017 at 14:33 Raffi Enficiaud notifications@github.com wrote:

This seems to be a Caffee problem. Which version of Caffee are you using? Is this the one that has been advertised on the main readme? ( a1c81aca641e5b16f3e2007be07dfdedc072606e)

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MPI-IS/bilateralNN/issues/6#issuecomment-302870450, or mute the thread https://github.com/notifications/unsubscribe-auth/APZ21zZNxAXWZFXo_VelP8ZR0_BAXBTUks5r7t2ygaJpZM4NgoEs .

[varunjampani]

Thanks for reporting the error. From the error, it seems there is some incompatibility with 'cudnn_tanh_layer'. Are you able to install Caffe without any bilateral (permutohedral) layers? Can you please try the manual way of adding the bilateralNN layers to your Caffe and try recompiling? Thanks.

[Sarah20187]

@raffienficiaud @varunjampani I tried the Manual way of Patching an existing Caffe version(reset the version to a1c81aca641e5b16f3e2007be07dfdedc072606e). It works. Thank you!

[varunjampani]

Thats great. Thanks for reporting. I am closing this issue then.

[hailingluo]

I have the similar issue when I make bilateralNN, the cudnn and CUDA parts always have errors. And I try the manual way of Patching and , but it still doesn't work . How to find the caffe version (a1c81aca641e5b16f3e2007be07dfdedc072606e) ? The caffe version which I am using is the one proposed by BLVC.

Errors: （part of) /home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_relu_layer.cu(18): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNReLULayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (52): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_relu_layer.cu(43): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNReLULayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (52): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_relu_layer.cu(18): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNReLULayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (52): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_relu_layer.cu(43): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNReLULayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (52): here

4 errors detected in the compilation of "/tmp/tmpxft_00005f34_00000000-7_cudnn_relu_layer.cpp1.ii". CMake Error at cuda_compile_generated_cudnn_relu_layer.cu.o.cmake:266 (message): Error generating file /home/lhl/bilateralNN/build/bin/src/caffe/CMakeFiles/cuda_compile.dir/layers/./cuda_compile_generated_cudnn_relu_layer.cu.o

src/caffe/CMakeFiles/caffe.dir/build.make:18511: recipe for target 'src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_cudnn_relu_layer.cu.o' failed make[5]: [src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_cudnn_relu_layer.cu.o] Error 1 make[5]: Waiting for unfinished jobs.... /home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_sigmoid_layer.cu(13): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNSigmoidLayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (42): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_sigmoid_layer.cu(33): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNSigmoidLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (42): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_sigmoid_layer.cu(13): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNSigmoidLayer::Forward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (42): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_sigmoid_layer.cu(33): error: argument of type "cudnnActivationMode_t" is incompatible with parameter of type "cudnnActivationDescriptor_t" detected during instantiation of "void caffe::CuDNNSigmoidLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (42): here

4 errors detected in the compilation of "/tmp/tmpxft_00005ecb_00000000-7_cudnn_sigmoid_layer.cpp1.ii". CMake Error at cuda_compile_generated_cudnn_sigmoid_layer.cu.o.cmake:266 (message): Error generating file /home/lhl/bilateralNN/build/bin/src/caffe/CMakeFiles/cuda_compile.dir/layers/./cuda_compile_generated_cudnn_sigmoid_layer.cu.o

src/caffe/CMakeFiles/caffe.dir/build.make:5908: recipe for target 'src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_cudnn_sigmoid_layer.cu.o' failed make[5]: *** [src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_cudnn_sigmoid_layer.cu.o] Error 1 /home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/include/caffe/util/cudnn.hpp(107): error: too few arguments in function call

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/include/caffe/util/cudnn.hpp(126): error: argument of type "int" is incompatible with parameter of type "cudnnNanPropagation_t"

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/include/caffe/util/cudnn.hpp(126): error: too few arguments in function call

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_conv_layer.cu(85): error: identifier "cudnnConvolutionBackwardFilter_v3" is undefined detected during instantiation of "void caffe::CuDNNConvolutionLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (123): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_conv_layer.cu(103): error: identifier "cudnnConvolutionBackwardData_v3" is undefined 3 errors detected during instantiation of "void caffe::CuDNNConvolutionLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=float]" (123): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_conv_layer.cu(85): error: identifier "cudnnConvolutionBackwardFilter_v3" is undefined detected in the compilation of "/tmp/tmpxft_00005dfa_00000000-5_base_data_layer.cpp4.ii". detected during instantiation of "void caffe::CuDNNConvolutionLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (123): here

/home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/src/caffe/layers/cudnn_conv_layer.cu(103): error: identifier "cudnnConvolutionBackwardData_v3" is undefined detected during instantiation of "void caffe::CuDNNConvolutionLayer::Backward_gpu(const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &, const std::vector<__nv_bool, std::allocator<__nv_bool>> &, const std::vector<caffe::Blob , std::allocator<caffe::Blob >> &) [with Dtype=double]" (123): here

4 errors detected in the compilation of "/tmp/tmpxft_00005f52_00000000-7_cudnn_conv_layer.cpp1.ii". /home/lhl/bilateralNN/build/tmp_caffe_clone/src/CaffeUpstream/include/caffe/util/cudnn.hpp(107): error: too few arguments in function call ...

src/caffe/CMakeFiles/caffe.dir/build.make:16734: recipe for target 'src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_pixel_feature_layer.cu.o' failed make[5]: [src/caffe/CMakeFiles/cuda_compile.dir/layers/cuda_compile_generated_pixel_feature_layer.cu.o] Error 1 CMakeFiles/Makefile2:272: recipe for target 'src/caffe/CMakeFiles/caffe.dir/all' failed make[4]: [src/caffe/CMakeFiles/caffe.dir/all] Error 2 Makefile:127: recipe for target 'all' failed make[3]: [all] Error 2 CMakeFiles/CaffeUpstream.dir/build.make:112: recipe for target 'tmp_caffe_clone/src/CaffeUpstream-stamp/CaffeUpstream-build' failed make[2]: [tmp_caffe_clone/src/CaffeUpstream-stamp/CaffeUpstream-build] Error 2 CMakeFiles/Makefile2:99: recipe for target 'CMakeFiles/CaffeUpstream.dir/all' failed make[1]: [CMakeFiles/CaffeUpstream.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: [all] Error 2

[varunjampani]

Seems like you have issues with CUDNN compatibility. Can you compile your default caffe without any of the bilateralNN layers? You can reset to the specific version with this command: git reset --hard a1c81aca641e5b16f3e2007be07dfdedc072606e or git checkout a1c81aca641e5b16f3e2007be07dfdedc072606e

[hailingluo]

@varunjampani Thanks for your help. But another new issue with CUDNN has arisen when I compile the caffe version (a1c81aca641e5b16f3e2007be07dfdedc072606e). I can compile my default caffe without any of the bilateralNN layers, but failed in the specific version. So I replace the old cudnn_* with the newest corresponding files to fix the problems. However when I compile the specific version caffe with bilateralNN, there are some errors with permutohedral_pooling_layer.

errors: src/caffe/layers/permutohedral_pooling_layer.cpp: In member function ‘virtual void caffe::PermutohedralPoolingLayer::LayerSetUp(const std::vector<caffe::Blob>&, const std::vector<caffe::Blob>&)’: src/caffe/layers/permutohedral_pooling_layer.cpp:22:3: error: ‘PermutohedralPoolingParameter’ was not declared in this scope PermutohedralPoolingParameter permutohedral_pooling_param = ^ src/caffe/layers/permutohedral_pooling_layer.cpp:25:23: error: ‘permutohedral_pooling_param’ was not declared in this scope do_repeatedinit = permutohedral_pooling_param.repeated_init(); ^ src/caffe/layers/permutohedral_pooling_layer.cpp: In member function ‘void caffe::PermutohedralPoolingLayer::InitLattice(const caffe::Blob, const caffe::Blob&, const caffe::Blob&, caffe::Blob)’: src/caffe/layers/permutohedral_pooling_layer.cpp:136:16: error: ‘PermutohedralPoolingParameter_NormType_AFTER’ was not declared in this scope case PermutohedralPoolingParameter_NormType_AFTER: ^ src/caffe/layers/permutohedral_pooling_layer.cpp:145:16: error: ‘PermutohedralPoolingParameter_NormType_SYMMETRIC’ was not declared in this scope case PermutohedralPoolingParameter_NormType_SYMMETRIC: ^ src/caffe/layers/permutohedral_pooling_layer.cpp: In instantiation of ‘void caffe::PermutohedralPoolingLayer::InitLattice(const caffe::Blob, const caffe::Blob&, const caffe::Blob&, caffe::Blob) [with Dtype = float]’: src/caffe/layers/permutohedral_pooling_layer.cpp:336:1: required from here src/caffe/layers/permutohedral_pooling_layer.cpp:135:11: error: ‘class caffe::LayerParameter’ has no member named ‘permutohedral_pooling_param’ switch (this->layerparam.permutohedral_pooling_param().norm_type()) { ^ src/caffe/layers/permutohedral_pooling_layer.cpp: In instantiation of ‘void caffe::PermutohedralPoolingLayer::InitLattice(const caffe::Blob, const caffe::Blob&, const caffe::Blob&, caffe::Blob) [with Dtype = double]’: src/caffe/layers/permutohedral_pooling_layer.cpp:336:1: required from here src/caffe/layers/permutohedral_pooling_layer.cpp:135:11: error: ‘class caffe::LayerParameter’ has no member named ‘permutohedral_pooling_param’ Makefile:572: recipe for target '.build_release/src/caffe/layers/permutohedral_pooling_layer.o' failed make: *** [.build_release/src/caffe/layers/permutohedral_pooling_layer.o] Error 1