Gold BlogHow to Setup a Python Environment for Machine Learning

In this tutorial, you will learn how to set up a stable Python Machine Learning development environment. You’ll be able to get right down into the ML and never have to worry about installing packages ever again.



Header image

Setting up your Python environment for Machine Learning can be a tricky task. If you’ve never set up something like that before, you might spend hours fiddling with different commands trying to get the thing to work. But we just want to get right to the ML!

In this tutorial, you will learn how to set up a stable Python Machine Learning development environment. You’ll be able to get right down into the ML and never have to worry about installing packages ever again.

 

(1) Set up Python 3 and Pip

 
The first step is to install pip , a Python package manager:

sudo apt-get install python3-pip


Using pip, we’ll be able to install any Python package that’s indexed in the Python Package Index with a simple pip install your_package . You’ll see soon how we use it to set up our virtual environment too.

Next, we’ll set Python 3 to be the default when running either the pip or python commands from command line. This makes using Python 3 easier and more convenient. If we didn’t do this, then if we wanted to use Python 3, we’d have to remember to type out pip3 and python3 every time!

To force Python 3 to be the default, we’re going to modify the ~/.bashrc file. From the command line, execute the following command to view that file:

nano ~/.bashrc


Scroll on down to the # some more ls aliases section and add the following line:

alias python='python3'


Save the file and reload your changes:

source ~/.bashrc


Boom! Python 3 is now your default Python! You can run it with a simple python your_program on the command line.

 

(2) Create a virtual environment

 
Now we’ll set up a virtual environment. In there, we’ll install all of the python packages that we need for Machine Learning.

We use virtual environments in order to separate our coding set ups. Imagine if at some point you wanted to do 2 different projects on your computer, which required different libraries of different versions. Having them all in the same working environment can be messy and you’ll likely run into the problem of conflicting library versions. Your ML code for project 1 needs version 1.0 of numpy, but project 2 needs version 1.15. Yikes!

A virtual environment allows us to isolate our working areas to avoid those conflicts.

First, install the relevant packages:

sudo pip install virtualenv virtualenvwrapper


Once we have virtualenv and virtualenvwrapper installed, we’ll again need to edit our ~/.bashrc file. Place these 3 lines right at the bottom and save it.

export WORKON_HOME=$HOME/.virtualenvs
export VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3
source /usr/local/bin/virtualenvwrapper.sh


Save the file and reload your changes:

source ~/.bashrc


Great! Now we can finally create our virtual environment like so:

mkvirtualenv ml


We’ve just created a virtual environment called ml . To enter it, do this:

workon ml


Nice! Any library installations that you do while in the ml virtualenv will be isolated in there and never conflict with any other environments! So whenever you wish to run code that depends on libraries installed in the mlenvironment, enter into first with the workon command and then run your code as normal.

If you need to exit the virtualenv, run this command:

deactivate

 

(3) Install Machine Learning libraries

 
Now we can install our ML libraries! We’ll go with the most commonly used ones:

  • numpy: for any work with matrices, especially math operations
  • scipy: scientific and technical computing
  • pandas: data handling, manipulation, and analysis
  • matplotlib: data visualisation
  • scikit learn: machine learning

Here’s a simple trick to install all of those libraries in one quick shot! Create a requirements.txt file and list all of the packages you wish to install like so:

numpy
scipy
pandas
matplotlib
scikit-learn


Once that’s done, just execute this command:

pip install -r requirements.txt


Voila! Pip will go ahead and install all of the packages listed in the file in one shot.

Congratulations, your environment is set up and you’re ready to do Machine Learning!

 

Like to learn?

 
Follow me on twitter where I post all about the latest and greatest AI, Technology, and Science!

 
Bio: George Seif is a Certified Nerd and AI / Machine Learning Engineer.

Original. Reposted with permission.

Related: