Incubator-mxnet: [lib/libmxnet.so] Error installing mxnet for EI Capitan 10.11.4

Created on 4 May 2016 · 13Comments · Source: apache/incubator-mxnet

My GPU: Intel HD Graphics 4000 + NVIDIA GeForce GT 650M

I have downgrade my XCode to V6.4 and my clang Version:

Apple LLVM version 6.1.0 (clang-602.0.51) (based on LLVM 3.6.0svn)
Target: x86_64-apple-darwin15.4.0
Thread model: posix

Now when I try to install Mxnet with either clang or clang-omp, I always get this error:

/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ranlib: file: lib/libmxnet.a(mxrtc.o) has no symbols
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ranlib: file: lib/libmxnet.a(cudnn_batch_norm.o) has no symbols
Undefined symbols for architecture x86_64:
      ...
  NOTE: a missing vtable usually means the first non-inline virtual member function has no definition.
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [lib/libmxnet.so] Error 1

Here is my osx.mk:

export CC = clang
export CXX = clang++
export NVCC = nvcc
DEBUG = 0
ADD_LDFLAGS = -L/usr/local/opt/openblas/lib
ADD_LDFLAGS += -Xlinker -F/Library/Frameworks -Xlinker -framework -Xlinker CUDA
ADD_CFLAGS = -I/usr/local/opt/openblas/include
USE_CUDA = 1
USE_CUDA_PATH = /usr/local/cuda
USE_CUDNN = 1
USE_NVRTC = 0
USE_OPENMP = 0
USE_BLAS = openblas
USE_INTEL_PATH = NONE

I checked my make version:

make -v
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for i386-apple-darwin11.3.0

Is i386 a problem?

Source

celestesosmart

Most helpful comment

Same issue for Mac El Capitan 10.11.5

tlkahn on 12 Jun 2016

👍2

All 13 comments

you might want to disable CUDNN and RTC, as they might not be supporting 32 bit?

tqchen on 7 May 2016

Still the same error, this seems not work for me:(

celestesosmart on 7 May 2016

One thing we might try, is to simply manually delete the cudnn_batchnorm.cu and mxrtc.cc and clean and recompile again

tqchen on 7 May 2016

Hi, thanks for your help. This still doesn't work for me. And here is a full version of error msg.

ld: symbol(s) not found for architecture x86_64
clang-3.5: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [lib/libmxnet.so] Error 1
make: *** Waiting for unfinished jobs....
In file included from tools/im2rec.cc:22:
In file included from /usr/local/Cellar/opencv/2.4.12_2/include/opencv2/opencv.hpp:56:
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:961:18: warning: 'CvForestTree::train'
      hides overloaded virtual functions [-Woverloaded-virtual]
    virtual bool train( CvDTreeTrainData* trainData, const CvMat* _subsample_idx, CvRTrees* forest );
                 ^
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:867:18: note: hidden overloaded virtual
      function 'CvDTree::train' declared here: different number of parameters (2 vs 3)
    virtual bool train( CvMLData* trainData, CvDTreeParams params=CvDTreeParams() );
                 ^
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:877:26: note: hidden overloaded virtual
      function 'CvDTree::train' declared here: different number of parameters (8 vs 3)
    CV_WRAP virtual bool train( const cv::Mat& trainData, int tflag,
                         ^
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:1149:18: warning: 'CvBoostTree::train'
      hides overloaded virtual functions [-Woverloaded-virtual]
    virtual bool train( CvDTreeTrainData* trainData,
                 ^
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:867:18: note: hidden overloaded virtual
      function 'CvDTree::train' declared here: different number of parameters (2 vs 3)
    virtual bool train( CvMLData* trainData, CvDTreeParams params=CvDTreeParams() );
                 ^
/usr/local/Cellar/opencv/2.4.12_2/include/opencv2/ml/ml.hpp:877:26: note: hidden overloaded virtual
      function 'CvDTree::train' declared here: different number of parameters (8 vs 3)
    CV_WRAP virtual bool train( const cv::Mat& trainData, int tflag,
                         ^
2 warnings generated.
Undefined symbols for architecture x86_64:
  "__ZNKSt13runtime_error4whatEv", referenced from:
      __ZTVN4dmlc5ErrorE in im2rec-5afd5e.o
      __ZTVN4dmlc5ErrorE in resource.o
      __ZTVN4dmlc10ParamErrorE in resource.o
      __ZTVN4dmlc5ErrorE in c_api.o
      __ZTVN4dmlc5ErrorE in c_predict_api.o
      __ZTVN4dmlc5ErrorE in engine.o
      __ZTVN4dmlc5ErrorE in naive_engine.o
      ...
  "__ZNKSt9exception4whatEv", referenced from:
      __ZTVNSt3__117bad_function_callE in c_api.o
      __ZTVNSt3__117bad_function_callE in naive_engine.o
      __ZTVNSt3__117bad_function_callE in threaded_engine.o
      __ZTVNSt3__117bad_function_callE in threaded_engine_perdevice.o
      __ZTVNSt3__117bad_function_callE in threaded_engine_pooled.o
      __ZTVNSt3__117bad_function_callE in image_aug_default.o
      __ZTVNSt3__117bad_function_callE in iter_csv.o
      ...
  "__ZNSt12length_errorD1Ev", referenced from:
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api_error.o
      ...
  "__ZNSt12out_of_rangeD1Ev", referenced from:
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in c_api.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in threaded_engine.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in threaded_engine_perdevice.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE15GetIORunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE13GetRunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in image_aug_default.o
      ...
  "__ZNSt13runtime_errorD2Ev", referenced from:
      __ZN4dmlc5ErrorD1Ev in im2rec-5afd5e.o
      __ZN4dmlc5ErrorD0Ev in im2rec-5afd5e.o
      __ZN4dmlc5ErrorD1Ev in resource.o
      __ZN4dmlc5ErrorD0Ev in resource.o
      __ZN4dmlc10ParamErrorD1Ev in resource.o
      __ZN4dmlc10ParamErrorD0Ev in resource.o
      __ZN4dmlc5ErrorD1Ev in c_api.o
      ...
  "__ZNSt9exceptionD2Ev", referenced from:
      __ZNSt3__117bad_function_callD1Ev in c_api.o
      __ZNSt3__117bad_function_callD0Ev in c_api.o
      __ZNSt3__117bad_function_callD1Ev in naive_engine.o
      __ZNSt3__117bad_function_callD0Ev in naive_engine.o
      __ZNSt3__117bad_function_callD1Ev in threaded_engine.o
      __ZNSt3__117bad_function_callD0Ev in threaded_engine.o
      __ZNSt3__117bad_function_callD1Ev in threaded_engine_perdevice.o
      ...
  "__ZSt9terminatev", referenced from:
      ___clang_call_terminate in im2rec-5afd5e.o
      ___clang_call_terminate in resource.o
      ___clang_call_terminate in c_api.o
      ___clang_call_terminate in c_predict_api.o
      ___clang_call_terminate in engine.o
      ___clang_call_terminate in naive_engine.o
      ___clang_call_terminate in threaded_engine.o
      ...
  "__ZTISt12length_error", referenced from:
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api_error.o
      ...
  "__ZTISt12out_of_range", referenced from:
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in c_api.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in threaded_engine.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in threaded_engine_perdevice.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE15GetIORunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE13GetRunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in image_aug_default.o
      ...
  "__ZTISt13runtime_error", referenced from:
      __ZTIN4dmlc5ErrorE in im2rec-5afd5e.o
      __ZTIN4dmlc5ErrorE in resource.o
      __ZTIN4dmlc5ErrorE in c_api.o
      __ZTIN4dmlc5ErrorE in c_predict_api.o
      __ZTIN4dmlc5ErrorE in engine.o
      __ZTIN4dmlc5ErrorE in naive_engine.o
      __ZTIN4dmlc5ErrorE in threaded_engine.o
      ...
  "__ZTISt9exception", referenced from:
      __ZTINSt3__117bad_function_callE in c_api.o
      __ZTINSt3__117bad_function_callE in naive_engine.o
      __ZTINSt3__117bad_function_callE in threaded_engine.o
      __ZTINSt3__117bad_function_callE in threaded_engine_perdevice.o
      __ZTINSt3__117bad_function_callE in threaded_engine_pooled.o
      __ZTINSt3__117bad_function_callE in image_aug_default.o
      __ZTINSt3__117bad_function_callE in iter_csv.o
      ...
  "__ZTVSt12length_error", referenced from:
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in im2rec-5afd5e.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api.o
      __ZNKSt3__121__basic_string_commonILb1EE20__throw_length_errorEv in c_api_error.o
      ...
  NOTE: a missing vtable usually means the first non-inline virtual member function has no definition.
  "__ZTVSt12out_of_range", referenced from:
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in resource.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in c_api.o
      __ZNKSt3__120__vector_base_commonILb1EE20__throw_out_of_rangeEv in threaded_engine.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in threaded_engine_perdevice.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE15GetIORunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZN5mxnet6engine13StreamManagerILm16ELm16EE13GetRunContextERKNS_7ContextE in threaded_engine_pooled.o
      __ZNK4dmlc9parameter10FieldEntryIiE10PrintValueERNSt3__113basic_ostreamIcNS3_11char_traitsIcEEEEi in image_aug_default.o
      ...
  NOTE: a missing vtable usually means the first non-inline virtual member function has no definition.
  "___kmpc_cancel_barrier", referenced from:
      _.omp_microtask. in libdmlc.a(data.o)
      _.omp_microtask.13 in libdmlc.a(data.o)
      _.omp_microtask.30 in libdmlc.a(data.o)
      _.omp_microtask.47 in libdmlc.a(data.o)
      _.omp_microtask.48 in libdmlc.a(data.o)
      _.omp_microtask.49 in libdmlc.a(data.o)
  "___kmpc_fork_call", referenced from:
      __ZN4dmlc4data18CreateLibSVMParserIjEEPNS_6ParserIT_EERKNSt3__112basic_stringIcNS6_11char_traitsIcEENS6_9allocatorIcEEEERKNS6_3mapISC_SC_NS6_4lessISC_EENSA_INS6_4pairISD_SC_EEEEEEjj in libdmlc.a(data.o)
      __ZN4dmlc4data18CreateLibSVMParserIyEEPNS_6ParserIT_EERKNSt3__112basic_stringIcNS6_11char_traitsIcEENS6_9allocatorIcEEEERKNS6_3mapISC_SC_NS6_4lessISC_EENSA_INS6_4pairISD_SC_EEEEEEjj in libdmlc.a(data.o)
      __ZN4dmlc4data9CSVParserIjEC2EPNS_10InputSplitERKNSt3__13mapINS5_12basic_stringIcNS5_11char_traitsIcEENS5_9allocatorIcEEEESC_NS5_4lessISC_EENSA_INS5_4pairIKSC_SC_EEEEEEi in libdmlc.a(data.o)
      __ZN4dmlc4data14TextParserBaseIjE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIjEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
      __ZN4dmlc4data14TextParserBaseIyE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIyEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
  "___kmpc_global_thread_num", referenced from:
      __ZN4dmlc4data14TextParserBaseIjE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIjEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
      __ZN4dmlc4data14TextParserBaseIyE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIyEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
  "___kmpc_push_num_threads", referenced from:
      __ZN4dmlc4data14TextParserBaseIjE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIjEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
      __ZN4dmlc4data14TextParserBaseIyE8FillDataEPNSt3__16vectorINS0_17RowBlockContainerIyEENS3_9allocatorIS6_EEEE in libdmlc.a(data.o)
  "_omp_get_num_procs", referenced from:
      _.omp_microtask.30 in libdmlc.a(data.o)
      _.omp_microtask.49 in libdmlc.a(data.o)
  "_omp_get_num_threads", referenced from:
      _.omp_microtask. in libdmlc.a(data.o)
      _.omp_microtask.47 in libdmlc.a(data.o)
  "_omp_get_thread_num", referenced from:
      _.omp_microtask.13 in libdmlc.a(data.o)
      _.omp_microtask.48 in libdmlc.a(data.o)
ld: symbol(s) not found for architecture x86_64
clang-3.5: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [bin/im2rec] Error 1

celestesosmart on 9 May 2016

Same issue for Mac El Capitan 10.11.5

tlkahn on 12 Jun 2016

👍2

@celestesosmart @tlkahn Have you disabled tried to disable OpenMP in dmlc-core? See https://github.com/dmlc/dmlc-core/pull/135

sxjscience on 12 Jun 2016

@sxjscience yes, we have disabled OpenMP by setting up config.mx. But still seeing errors in linking stage... My log (part on errors) is as follows:

Undefined symbols for architecture x86_64:
"_cblas_dgemm", referenced from:
mxnet::op::ConvolutionOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in convolution.o
mxnet::op::ConvolutionOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in convolution.o
mxnet::op::DeconvolutionOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in deconvolution.o
mxnet::op::DeconvolutionOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in deconvolution.o
mxnet::op::FullyConnectedOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in fully_connected.o
mxnet::op::FullyConnectedOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in fully_connected.o
"_cblas_sdot", referenced from:
void mshadow::VectorDot(mshadow::Tensor, mshadow::Tensor const&, mshadow::Tensor const&) in broadcast_reduce_op.o
void mshadow::VectorDot(mshadow::Tensor, mshadow::Tensor const&, mshadow::Tensor const&) in matrix_op.o
"_cblas_sgemm", referenced from:
mxnet::op::ConvolutionOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in convolution.o
mxnet::op::ConvolutionOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in convolution.o
mxnet::op::DeconvolutionOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in deconvolution.o
mxnet::op::DeconvolutionOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in deconvolution.o
mxnet::op::FullyConnectedOp::Forward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in fully_connected.o
mxnet::op::FullyConnectedOp::Backward(mxnet::OpContext const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&, std::__1::vector > const&) in fully_connected.o
void mxnet::op::DotForward_mshadow::cpu(mshadow::TBlob const&, mshadow::TBlob const&, mxnet::op::EnvArguments const&, mshadow::TBlob_, mxnet::OpReqType, mxnet::RunContext) in matrix_op.o
...
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *_* [bin/im2rec] Error 1

shi27feng on 12 Jun 2016

@tqchen Please help, we are stuck (and bleeding).......

shi27feng on 12 Jun 2016

👍1

@shi27feng, According to the error message, the problem is in linking blas libraries. Also, on OSX, directly setting the USE_OPENMP flag in config.mk will not disable OpenMP completely, you'd better apply the fix in https://github.com/dmlc/dmlc-core/pull/135

sxjscience on 13 Jun 2016

@shi27feng Do you know how to fix this issue now? I have the same problem on make

rehanrishi21 on 26 Oct 2016

I ran into the same issue. I set USE_OPENMP = 1 and it's compiled.

wddabc on 8 Mar 2017

I ran into the same issue with Macbook Pro late 2015, 10.12.3 Sierra, gcc-7.

I believe that the issue is cause by two versions of OpenCV (2.4.13 and 3.2), because I ran into OpenCV issue several times on my MBP.
So, I tried 2.4.13.2 and 3.20 and HEAD-3649ee3, and only 2.4.13 works, with g++/clang++, not g++-7.

mxnet daodao$ g++ --version
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 8.1.0 (clang-802.0.42)
Target: x86_64-apple-darwin16.4.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin

The most important thing here is to setup OpenCV version explicitly:

# setup opencv
ifeq ($(USE_OPENCV), 1)
    CFLAGS  += -DMXNET_USE_OPENCV=1 -I/usr/local/Cellar/opencv/2.4.13.2/include -I/usr/local/include
    LDFLAGS += $(filter-out -lopencv_ts, -L/usr/local/Cellar/opencv/2.4.13.2/lib -lopencv_shape -lopencv_stitching -lopencv_objdetect -lopencv_superres -lopencv_videostab -lopencv_calib3d -lopencv_features2d -lopencv_highgui -lopencv_videoio -lopencv_imgcodecs -lopencv_video -lopencv_photo -lopencv_ml -lopencv_imgproc -lopencv_flann -lopencv_core)
    #CFLAGS += -DMXNET_USE_OPENCV=1 $(shell pkg-config --cflags opencv)
    #LDFLAGS += $(filter-out -lopencv_ts, $(shell pkg-config --libs opencv))
    BIN += bin/im2rec
else
    CFLAGS+= -DMXNET_USE_OPENCV=0
endif

zzningxp on 1 Sep 2017

Mac OSX python users have the option to use pre-built pip packages now. The package names are mxnet and mxnet-mkl. We also offer nightly builds as beta pre-release, available through pip's --pre switch. Starting from 0.11.0, the pip packages depend on opencv 3.3. Though a hack is needed to use the compatible header in imgcodecs.

Feel free to ping me to reopen if there's still issue in building OSX version.