Tensor flow install OSX

Question

I'm trying to install TensorFlow on El 10.11.1, and I'm having issues creating and installing the pip package that I do not understand at all.

I successfully installed Bazel, SWIG and NumPy per instructions (at least they all seem to work at this time).

Creating the pip package/install I get the following error when I run the command:

bazel build -c opt //tensorflow/tools/pip_package:build_pip_package

/Users/christophermichael/tensorflow/tensorflow/tensorflow/python/BUILD:730:1: C++ compilation of rule '//tensorflow/python:_pywrap_tensorflow.so' failed: osx_gcc_wrapper.sh failed: error executing command

Running with --verbose_failures, I get:

/Users/christophermichael/tensorflow/tensorflow/tensorflow/python/BUILD:730:1: C++ compilation of rule '//tensorflow/python:_pywrap_tensorflow.so' failed: osx_gcc_wrapper.sh failed: error executing command 
(cd /private/var/tmp/_bazel_christophermichael/b14fc0915a4b8dd526e88d50650b0640/tensorflow && \
  exec env - \
    INTERCEPT_LOCALLY_EXECUTABLE=1 \
    PATH=/usr/local/bin:/usr/local/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/opt/X11/bin:/Users/christophermichael/bin \
    TMPDIR=/var/folders/tv/rgqv4fp54mn952pn_v0tjwt80000gn/T/ \
  tools/cpp/osx_gcc_wrapper.sh '-D_FORTIFY_SOURCE=1' -fstack-protector -fcolor-diagnostics -Wall -Wthread-safety -Wself-assign -fno-omit-frame-pointer -g0 -O2 -DNDEBUG -ffunction-sections -fdata-sections '-std=c++0x' -iquote . -iquote bazel-out/local_darwin-opt/genfiles -isystem google/protobuf/src -isystem bazel-out/local_darwin-opt/genfiles/google/protobuf/src -isystem tools/cpp/gcc3 -isystem external/jpeg_archive/jpeg-9a -isystem bazel-out/local_darwin-opt/genfiles/external/jpeg_archive/jpeg-9a -isystem external/png_archive/libpng-1.2.53 -isystem bazel-out/local_darwin-opt/genfiles/external/png_archive/libpng-1.2.53 -isystem external/re2 -isystem bazel-out/local_darwin-opt/genfiles/external/re2 -isystem third_party/gpus/cuda -isystem bazel-out/local_darwin-opt/genfiles/third_party/gpus/cuda -isystem third_party/gpus/cuda/include -isystem bazel-out/local_darwin-opt/genfiles/third_party/gpus/cuda/include -isystem third_party/eigen3 -isystem bazel-out/local_darwin-opt/genfiles/third_party/eigen3 -I/System/Library/Frameworks/Python.framework/Versions/2.7/Extras/lib/python/numpy/core/include -Wno-self-assign -Wno-write-strings -I/usr/include/python2.7 -no-canonical-prefixes -Wno-builtin-macro-redefined '-D__DATE__="redacted"' '-D__TIMESTAMP__="redacted"' '-D__TIME__="redacted"' '-frandom-seed=bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.o' -MD -MF bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.d -fPIC -c bazel-out/local_darwin-opt/bin/tensorflow/python/pywrap_tensorflow.cc -o bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.o): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 1: osx_gcc_wrapper.sh failed: error executing command 
  (cd /private/var/tmp/_bazel_christophermichael/b14fc0915a4b8dd526e88d50650b0640/tensorflow && \
  exec env - \
    INTERCEPT_LOCALLY_EXECUTABLE=1 \
    PATH=/usr/local/bin:/usr/local/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/opt/X11/bin:/Users/christophermichael/bin \
    TMPDIR=/var/folders/tv/rgqv4fp54mn952pn_v0tjwt80000gn/T/ \
  tools/cpp/osx_gcc_wrapper.sh '-D_FORTIFY_SOURCE=1' -fstack-protector -fcolor-diagnostics -Wall -Wthread-safety -Wself-assign -fno-omit-frame-pointer -g0 -O2 -DNDEBUG -ffunction-sections -fdata-sections '-std=c++0x' -iquote . -iquote bazel-out/local_darwin-opt/genfiles -isystem google/protobuf/src -isystem bazel-out/local_darwin-opt/genfiles/google/protobuf/src -isystem tools/cpp/gcc3 -isystem external/jpeg_archive/jpeg-9a -isystem bazel-out/local_darwin-opt/genfiles/external/jpeg_archive/jpeg-9a -isystem external/png_archive/libpng-1.2.53 -isystem bazel-out/local_darwin-opt/genfiles/external/png_archive/libpng-1.2.53 -isystem external/re2 -isystem bazel-out/local_darwin-opt/genfiles/external/re2 -isystem third_party/gpus/cuda -isystem bazel-out/local_darwin-opt/genfiles/third_party/gpus/cuda -isystem third_party/gpus/cuda/include -isystem bazel-out/local_darwin-opt/genfiles/third_party/gpus/cuda/include -isystem third_party/eigen3 -isystem bazel-out/local_darwin-opt/genfiles/third_party/eigen3 -I/System/Library/Frameworks/Python.framework/Versions/2.7/Extras/lib/python/numpy/core/include -Wno-self-assign -Wno-write-strings -I/usr/include/python2.7 -no-canonical-prefixes -Wno-builtin-macro-redefined '-D__DATE__="redacted"' '-D__TIMESTAMP__="redacted"' '-D__TIME__="redacted"' '-frandom-seed=bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.o' -MD -MF bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.d -fPIC -c bazel-out/local_darwin-opt/bin/tensorflow/python/pywrap_tensorflow.cc -o bazel-out/local_darwin-opt/bin/tensorflow/python/_objs/_pywrap_tensorflow.so/tensorflow/python/pywrap_tensorflow.pic.o): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 1.
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 4.940s, Critical Path: 3.69s

Answer 1

Yeah, as Hamed said above, it's better directly install the binaries.

However, if you do need to build from the source, then these steps should work:

• brew install python
• virtualenv ~/tf
• source ~/tf/bin/activate
• Install bazel for Mac
• Install PCRE (not PCRE2) for Mac
• Install SWIG for Mac
• git clone --recurse-submodules https://github.com/tensorflow/tensorflow
• cd tensorflow
• ./configure
• bazel build -c opt //tensorflow/tools/pip_package:build_pip_package
• bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
• pip install /tmp/tensorflow_pkg/tensorflow-0.5.0-cp27-none-linux_x86_64.whl
• bazel build -c opt //tensorflow/cc:tutorials_example_trainer

Answer 2

Another option is to install anaconda and use its environment management. The advantage of using environments either with anaconda or virtualenv is that you separate sources and dependencies for different use cases. This means, if you need to make changes to your python libraries for something else, it won't affect your tensorflow environmnet.

Here is how to install tensorflow with anaconda.

1. Install anaconda http://docs.continuum.io/anaconda/install
2. Create a new conda workspace
3. Download the specific protobuf version that tensorflow needs: https://pypi.python.org/pypi/protobuf/3.0.0a3
4. Install it via sudo easy_install ~/Downloads/protobuf-3.0.0a3-py2.7.egg
5. Install a numpy version greater than 1.08.x via conda install numpy

6. Download the 0.6.0 version of tensorflow: https://storage.googleapis.com/tensorflow/mac/tensorflow-0.6.0-py2-none-any.whl

7. Install via pip install ~/Downloads/tensorflow-0.6.0-py2-none-any.whl

When you install tensorflow from the whl file directly, it should tell you when dependencies are not there. It seems not to be able to resolve these conflicts independently. My setup had issues with protobuf and numpy. After installing them manually everything worked fine.

I hope this helps!