Tensorflow CPU version Installation on Ubuntu 16.04

1. Create new user with administration privilege

Add new user

adduser XXXX

Add administrator privilege

visudo

Add following line into file

XXXX ALL=(ALL:ALL) ALL

2.Install Anaconda

Download and Installation

wget https://repo.anaconda.com/archive/Anaconda3-5.2.0-Linux-x86_64.sh
sh Anaconda3-5.2.0-Linux-x86_64.sh

Reload PATH setting

source .bashrc

Test

python

Python 3.6.5 |Anaconda, Inc.| (default, Apr 29 2018, 16:14:56)
[GCC 7.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.

Python 3.6.5 install successfully

3. Install Other Libraries

pip install msgpack
pip install xgboost

4.Install Tensorflow

Create dedicate environment

conda create -n tensorflow
source activate tensorflow

Install Tensorflow ready compiled package

pip install --ignore-installed --upgrade https://files.pythonhosted.org/packages/04/79/a37d0b373757b4d283c674a64127bd8864d69f881c639b1ee5953e2d9301/tensorflow-1.5.0-cp36-cp36m-manylinux1_x86_64.whl

Test

python

import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!') 
sess = tf.Session() 
print(sess.run(hello)) 

h5py future warning

import tensorflow as tf
/home/dzhu/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters

Fix: Should ypgrade h5py to version 2.8.0
pip install h5py==2.8.0

Issues

2018-06-16 14:09:21.414634: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2

因为通过pip安装的是manylinux1_X86这是一个通用版本，而目前CPU所支持SSE2个指令集，通用版本并没有用到.
如果要解决这个问题，需要对Tensorflow源码进行不同的编译再安装。