Install Tensorflow on Ubuntu (17.10)

Install Tensorflow on Ubuntu (17.10)

Requirements

• An NVIDIA GPU with a compute capability of 3.0 or higher.
  • I'll be using a GTX 1050 Ti hooked up externally to a thinkpad T420s through the express card port
• Python installed.
  • Ubuntu comes with python already installed but I will be using anaconda

Overview

• Step 1: Update your GPU driver (should be higher than version 390)
• Step 2: Install the CUDA Toolkit version 9.0 (with all the patches)
• Step 3: Install CUDNN 7.0.5
• Step 4: Install Tensorflow GPU with pip
• Step 5: Test it!

Step 1: Update your GPU driver

Open a terminal and run the following 3 commands

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
sudo apt install nvidia-390

Reboot your computer. To verify the installation, open a terminal and run the following command

nvidia-smi

The output should show the GPU name and the driver

Step 2: Install the CUDA Toolkit (9.0)

• go to https://developer.nvidia.com/cuda-90-download-archive and download the toolkit for linux, x86_64, ubuntu, 17.04, deb l
• once the download is complete, open a terminal in the directory the base installer is and run the follow commands

sudo dpkg -i cuda-repo-ubuntu1704-9-0-local_9.0.176-1_amd64.deb
sudo apt-key add /var/cuda-repo-9-0-local/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda

• download patch 1 and install (you should get a prompt to install once its done downloading)
• download patch 2 and install (you should get a prompt to install once its done downloading)
• open your .bashrc file with nano

sudo nano ~/.bashrc

• go to the last line and add the following lines (this will set your PATH variable)

export PATH=/usr/local/cuda-9.0/bin${PATH:+:$PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-9.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Step 3: Install CUDNN 7.0.5

• go to https://developer.nvidia.com/cudnn
• Select CUDNN 7.0.5 for CUDA 9.0
• download the cuDNN v7.0.5 Library for Linux (tar file)
• open a terminal in the directory the tar file is located
• unzip the tar file using the command

tar -xzvf cudnn-9.0-linux-x64-v7.tgz

• run the following commands to move the appropriate files to the CUDA folder

sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

Step 4: pip install tensorflow-gpu

I will be using a conda environment for installing tensorflow

• create a conda environment by using the following command

conda create -n tf python=3.6 pip

• activate your environment using

source activate tf

run the following command to install tensorflow

pip install tensorflow-gpu==1.5

Step 5: Test it!

• start a python interpreter in the terminal by typing

python

• run the following lines

>>> import tensorflow as tf
>>> hello = tf.constant('hello tensorflow')
>>> with tf.Session() as sesh:
>>>     sesh.run(hello)

• the output should be

>>> 'hello tensorflow'