Tensorflow源码安装

以Ubuntu 16.04为例,介绍Tensorflow源码安装的方法。

下载tensorflow源码

git clone https://github.com/tensorflow/tensorflow

安装bazel

echo "deb [arch=amd64] http://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
curl https://bazel.build/bazel-release.pub.gpg | sudo apt-key add -
sudo apt-get update && sudo apt-get install bazel

安装依赖库

# Python 2.7
sudo apt-get install python-numpy python-dev python-pip python-wheel
# Python 3.x
sudo apt-get install python3-numpy python3-dev python3-pip python3-wheel

安装CUDA和cuDNN

参考这里

编译安装

cd tensorflow
./configure

安装命令行提示,逐个设置编译选项(可以选择默认值)。

编译CPU版:

bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package

编译GPU版:

bazel build --config=opt --config=cuda //tensorflow/tools/pip_package:build_pip_package 

注意,GCC 5需要设置--cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0"选项。

bazel build会生成一个build_pip_package命令,用来生成python whl包:

# 编译生成python whl包
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

最后,安装生成的包

sudo pip install /tmp/tensorflow_pkg/tensorflow-1.2.0-py2-none-any.whl