Install Tensorflow on Linux Ubuntu 18.04 LTS

Step 1: Update your GPU drivers

The GPU driver should be higher than version 390. You can check what graphics driver you have installed with the following command:

nvidia-smi

The output should be like the following:

Wed Aug 22 19:25:07 2018 
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.48                 Driver Version: 390.48                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 950M    Off  | 00000000:01:00.0 Off |                  N/A |
| N/A   53C    P0    N/A /  N/A |    237MiB /  4046MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      1071      G   /usr/lib/xorg/Xorg                            24MiB |
|    0      1165      G   /usr/bin/gnome-shell                          48MiB |
|    0      1561      G   /usr/lib/xorg/Xorg                            93MiB |
|    0      1747      G   /usr/bin/gnome-shell                          66MiB |
+-----------------------------------------------------------------------------+

Step 2: Install the CUDA Toolkit 9.0

Go to https://developer.nvidia.com/cuda-90-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1704&target_type=runfilelocal and download the toolkit for Linux - x86_64 - Ubuntu - 17.04 - runfile (local) - Base installer.

Once you’ve got that file, navigate to where the file was downloaded. Open a terminal and run the following commands:

sudo chmod +x cuda_9.0.176_384.81_linux.run
./cuda_9.0.176_384.81_linux.run --override

Note

Accept the term and conditions.

Do you accept the previously read EULA?
accept/decline/quit: accept

Say yes to installing with an unsupported configuration.

You are attempting to install on an unsupported configuration. Do you wish to continue?
(y)es/(n)o [ default is no ]: yes

Say no to installing NVIDIA Accelerated Graphics Driver.

Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 384.81?
(y)es/(n)o/(q)uit: no

Say yes to install the CUDA 9.0 Toolkit.

Install the CUDA 9.0 Toolkit?
(y)es/(n)o/(q)uit: yes

Add these lines to the end of ~/.bashrc:

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

Restart the terminal.

Step 3: Install cuDNN 7.0.5

Download https://developer.nvidia.com/compute/machine-learning/cudnn/secure/v7.0.5/prod/9.0_20171129/cudnn-9.0-linux-x64-v7

Once you’ve got that file, navigate to where the file was downloaded. Open a terminal and run the following commands:

tar -zxvf cudnn-9.0-linux-x64-v7.tgz
sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda-9.0/lib64/
sudo cp  cuda/include/cudnn.h /usr/local/cuda-9.0/include/
sudo chmod a+r /usr/local/cuda-9.0/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

Step 4: Install Anaconda

Download Anaconda for Linux: https://www.anaconda.com/download/#linux

Run the Anaconda script:

sh Anaconda3-5.2.0-Linux-x86_64.sh

Create an Anaconda environment named tf with Python 3.6:

conda create -n tf pip python=3.6

Activate the installation with the following command:

source activate tf

Step 5: Instal TensorFlow GPU

Run:

pip install tensorflow-gpu==1.5

(Optional) Step 6: Install Keras