How to Setup a Python Environment for Machine Learning on Linux or Ubuntu Machine

This tutorial covers a short and straight forward step by step procedure required to set up a flexible machine learning environment on a Linux/ubuntu machine using Anaconda.

Installing Anaconda

1. Download the Anaconda package from the Anaconda official website

cd /tmp 
curl -O https://repo.anaconda.com/archive/Anaconda3-2019.10-Linux-x86_64.sh

Remember, you can change the <2019.10> to the latest version number

2. To check the integrity of the installer

sha256sum Anaconda3-2019.10-Linux-x86_64.sh

3. Install Anaconda by running the downloaded installer script

bash Anaconda3-2019.10-Linux-x86_64.sh

Save(write down) the displayed directory: we will use it later

Follow the below steps to complete the installation

1. Press Enter
2. Type “yes”
3. Confirm the installation location
4. Initialize Anaconda by answering “yes” when it gets to the following step

If you forget to say yes to the step shown below , use the code below to modify your shell scripts

eval "$(/home/cc/anaconda3/bin/conda shell.bash hook)"

4. Acitavate installation

source ~/.bashrc

You successfully installed and activated Anaconda

Test installation

Test the installation by listing the installed packages

conda list

Creating Anaconda Environment for Machine Learning Projects

In this section, we will create an environment called “torch_env” that will hold various machine learning libraries and packages.

1. Creating the Anaconda environment and we name it “torch_env”. You can give any other name

conda create -n torch_env python=3.6

2. Activate the newly created environment: Once it finishes creating the environment activate it as follows

conda activate torch_env

Don’t forget to always activate your environment before you install anything or start to work on your project

Now we are in the environment, so let’s start installing important machine learning libraries and frameworks. The environment helps to structure your work, to easily share it, to easily replicate it.

• Install pip to your environment directory

conda install pip

Now we will install the machine learning libraries! We’ll go with the most commonly used ones:

• NumPy: for any work with matrices, especially math operations
• pandas: data handling, manipulation, and analysis
• Pytorch: Deep learning framework

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. Your requirements file looks like this

Write down the following libraries and save them as requirements.txt (use this link to learn how to use vim editor for Linux)

vim requirements.txt

requirements.txt

Then install the requirements file

python3 -m pip install -r  requirements.txt

To verify you have installed the packages

• Open Python

python

• Import the libraries