Installing Keras Using Python & TensorFlow: A Guide

What is Keras?

Keras is a tool for machine learning specialists who work with Python, mostly used due to the convenience of mathematical calculations. Developers use Keras to create, configure, and test machine learning and artificial intelligence systems, primarily neural networks.

What are Keras Models?

Keras works with models or schemes by which information is distributed and transformed. Machine learning is processing information using a programmed network, where certain conclusions are drawn based on certain data. The network structure is called a model and is often presented as a graph, diagram, or table.

Prerequisites for Installing Keras

• A server with root-level access
• Python installed
• TensorFlow installed

This tutorial is performed with CentOS 7.

Install Keras via Python & TensorFlow

Install Python 3

To install Keras, Python is required to be installed on your computer since Keras is based on Python. It is recommended to have the latest version of Python, such as Python 3 for example.

Step 1: Update the Environment

In order to make sure that we are working with the most up-to-date environment possible in terms of our packages, we can run the following command:

[root@centos7 ~]# yum update -y

Step 2: Install Python 3 on CentOS 7

Now that the environment is up to date, all we need to do to install Python 3 is run the following command:

[root@centos7 ~]# yum install -y python3

That’s it! Python 3 is now installed! Another helpful idea to consider is that Pip (also know as Pip3), the Python package manager for Python 3, is installed alongside the Python 3 package, so we don’t have to worry about that as an additional installation step.

Step 3: Verify Installation

To ensure that Python 3 is installed and usable, we can drop it into a Python 3 shell by running the following command:

[root@centos7 ~]# python3
Python 3.6.8 (default, Aug  7 2019, 17:28:10) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-39)] on linux
Type "help", "copyright", "credits" or "license" for more information.

>>>

You should see the version of Python 3 installed on your system as well as a change in the command prompt characters.

However, in some cases, you might want to have the most recent version of Python available, and that’s where a source installation can come in handy for which we have a detailed article that covers installing Python 3 on CentOS 7.

Install TensorFlow

TensorFlow is one of the backend engines that we need to install before Keras can be installed.

Step 1: Create a Virtual Environment

You need to create a virtual environment when installing TensorFlow. First, create a folder for your project using the following commands:

# mkdir test
# cd test

Use the following command to create a virtual environment with Python. It creates a virtual environment named tf-virtual-env. Replace this with your chosen name:

# python3 -m venv tf-virtual-env

Inside this environment are many pre-installed Python libraries and tools needed in the project, such as the Package installer for Python (Pip). The following command activates the environment:

# source tf-virtual-env/bin/activate

The prompt changes in the terminal, which means you activated it successfully:

(tf-virtual-env) # _

Type deactivate at the prompt and press Enter to exit the environment.

TensorFlow requires at least version 19.0 of Pip. Use the following command to ensure you have the latest version:

# pip install --upgrade pip

Step 2: Installing TensorFlow

This tutorial installs a version that does not use your GPU. You can install a TensorFlow version that offers GPU support.

Use the following command to install TensorFlow without GPU support:

# pip install --upgrade tensorflow

Use the same command for updating TensorFlow.

Step 3: Test the Environment

To test your environment, open Python bash:

# python

Then, type the following command. The first line will import the TensorFlow packages into the Python interpreter session, while the second line will print the TensorFlow version:

import tensorflow as tf

 print(tf.__version__)'

Install Keras

Thanks to a new update in TensorFlow 2.0+, if you installed TensorFlow as instructed, you don’t need to install Keras anymore because it is installed with TensorFlow.

To confirm it, open Python bash:

# python

At the prompt, run the following commands:

import keras

keras.__version__

For those using TensorFlow versions before 2.0, here are the instructions for installing Keras using Pip.

Step 1: Installing Keras

Install Keras with the following command:

pip3 install keras

The terminal shows the confirmation message once the process completes:

Collecting keras
  Obtaining dependency information for keras from https://files.pythonhosted.org/packages/2e/f3/19da7511b45e80216cbbd9467137b2d28919c58ba1ccb971435cb631e470/keras-2.13.1-py3-none-any.whl.metadata
  Downloading keras-2.13.1-py3-none-any.whl.metadata (2.4 kB)
Downloading keras-2.13.1-py3-none-any.whl (1.7 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.7/1.7 MB 11.1 MB/s eta 0:00:00
Installing collected packages: keras
Successfully installed keras-2.13.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv

Step 2: Verify Installation of Keras

Verify the install of Keras by displaying the package information:

pip3 show keras

The output will be as shown below:

Name: keras
Version: 2.13.1
Summary: Deep learning for humans.
Home-page: https://keras.io/
Author: Keras team
Author-email: keras-users@googlegroups.com
License: Apache 2.0
Location: /opt/rh/rh-python38/root/usr/local/lib/python3.8/site-packages
Requires: 
Required-by: 

Deprecation of the Git Clone Keras Install Method

Keras was previously installed by cloning the GitHub repository, unpacking the packages, and installing the software. The Keras team deprecated the GitHub repository and moved the applications into the core Keras repository and the TensorFlow Pip package:

The recommended Keras installation method from the Keras team is via TensorFlow version 2+.

Importing Keras Libraries

When we create a Python virtual environment, it already contains most of the important libraries. Here is the list of the most important Python libraries:

• Numpy
• Pandas
• Scikit-learn
• Matplotlib
• Scipy
• Seaborn

To confirm the Numpy library, use the following command:

import numpy as np

ndarray = np.ones([3, 3])

print("TensorFlow operations convert numpy arrays to Tensors automatically")
tensor = tf.math.multiply(ndarray, 42)
print(tensor)

The output should be as follows:

tf.Tensor( [[42. 42. 42.] 
     [42. 42. 42.]
     [42. 42. 42.]], shape=(3, 3), dtype=float64)

Updating Keras

Since Keras installs alongside Tensorflow, they also update together. Use the command below, the same one to update TensorFlow, to update Keras:

# pip install --upgrade tensorflow

Conclusion

There are ways to install Karas and Tensorflow without a virtual environment. Still, it can be risky and more complex than the commands of pre-configured environments. Whether installing Keras using Pip via Python or TensorFlow, this tutorial helps you get it up and running for your next deep learning project.

Do you need a CentOS or AlmaLinux machine for your next Keras install? Liquid Web has these options for VPS Hosting, Cloud Dedicated Servers, and Dedicated Servers. Contact our sales team today for setup options.

Original Publication Date

This article was originally published in June 2022. It has since been updated for accuracy and comprehensiveness.