This repository contains a custom builds of tensorflow. To install one of these on your system, make sure you choose the correct file according to your version of python and gcc and run the following command pip install --ignore-installed --upgrade /path/to/binary.whl * tensorflow-1.1.0-cp36-cp36m-linux_x86_64.whl Flags: -msse4.1 -msse4.2 -mavx -mavx2 -mfma Other: Compiled with support for XLA JIT Config: Arch Linux, GCC 6.3, Python 3.6.1 * tensorflow-1.2.0rc1-cp36-cp36m-linux_x86_64.whl Flags: -march=native -mtune=native -O3 CPU Info: Intel(R) Core(TM) i7-4500U CPU @ 1.80GHz Config: Arch Linux, GCC 6.3, Python 3.6.1 * tensorflow-1.2.0rc1-cp35-cp35m-linux_x86_64.whl Flags: -msse4.1 -msse4.2 -mavx -mavx2 -mfma -O3 Config: Ubuntu 16.04, GCC 5.4, Python 3.5.2 * tensorflow-1.2.0rc2-cp36-cp36m-linux_x86_64.whl Flags: -march=native -mtune=native -O3 CPU Info: Intel(R) Core(TM) i7-4500U CPU @ 1.80GHz Config: Arch Linux, GCC 7.1, Python 3.6.1 Comments: Built with MKL support