Install Tensorflow 2.0 with GPU Support and Jupyter Notebook

Let’s begin.

Step 1: Add NVIDIA package repositories

# create temp folder 
mkdir ~/tmp && cd ~/tmp
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo apt-get update
wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt-get update

Step 2: Install the NVIDIA driver

sudo apt-get install --no-install-recommends nvidia-driver-418

After that, reboot the system and type the command below:

nvidia-smi

You should see GPUs on the system:

Step 3: Install development and runtime libraries

sudo apt-get install --no-install-recommends cuda-10-0 libcudnn7=7.6.2.24-1+cuda10.0 libcudnn7-dev=7.6.2.24-1+cuda10.0

Step 4 (Optional): Install TensorRT

TensorRT to improve latency and throughput for inference on some models.

sudo apt-get install -y --no-install-recommends libnvinfer5=5.1.5-1+cuda10.0 libnvinfer-dev=5.1.5-1+cuda10.0

Step 5 : Install Anaconda

cd ~/tmp
wget https://repo.anaconda.com/archive/Anaconda3-2019.07-Linux-x86_64.sh
bash ./Anaconda3-2019.07-Linux-x86_64.sh
# 1. Review the license agreement then type ENTER
# 2. For accept the license terms: type 'yes'
# 3. Choose installation location, default is '~/anaconda3'
# 4. For initialize Anaconda3, type 'yes'
source ~/.bashrc

After the installation is complete, you can prevent automatically activating the base anaconda environment with the following command:

conda config --set auto_activate_base false
source ~/.bashrc

Check python and pip location:

which python && which pip

This command should give ~/anaconda3/bin/python and ~/anaconda3/bin/pip locations.

Step 6: Install Jupyer Notebook with conda

conda install -c anaconda jupyter

Step 7 (Optional): Jupyter Notebook Access Remotely

By default, a notebook server runs locally at 127.0.0.1:8888 and is accessible only from localhost. You may access the notebook server from the browser using http://127.0.0.1:8888.

If you are working on a cloud platform (GCloud, AWS, etc.), you may want to access jupyter notebook remotely. The next step shows you how to configure jupyter notebook for accessing remotely.

7.1: Generate config file

The following command will generate the configuration file at ~/.jupyter/jupyter_notebook_config.py.

jupyter notebook --generate-config

7.2: Generate SSL Certificate and Key File

A self-signed certificate can be generated with openssl for encrypted communication.

mkdir ~/ssl
openssl req -x509 -nodes -days 365 -newkey rsa:2048 -keyout ~/ssl/key.key -out ~/ssl/cert.pem

7.3: Edit config file

Open the notebook config file (~/.jupyter/jupyter_notebook_config.py) with your favorite edit app like nano, vim and add this section top of the file:

c.NotebookApp.certfile = '/home/{USER}/ssl/cert.pem' # full path of cert.pem
c.NotebookApp.keyfile = '/home/{USER}/ssl/key.key'  # full path of cert.pem
c.NotebookApp.ip = '*'
c.NotebookApp.open_browser = False

7.4: Setting Password

jupyter notebook password

Step 8: Install Tensorflow 2.0 with pip

pip install tensorflow-gpu==2.0.0

Final Checks:

After these steps finally, you can start jupyter notebook with the following command:

jupyter notebook

Then open a new jupyter notebook file, and write these three lines of code.

from __future__ import absolute_import, division, print_function, unicode_literals
import tensorflow as tf
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))

You should see the count of available GPUs on the system.

Cheers!

References:

1. https://www.tensorflow.org/guide/gpu
2. https://jupyter-notebook.readthedocs.io/en/stable/public_server.html