Minimalistic, yet productive, Machine Learning Platform with Miniconda and Python

Setting up a robust Machine Learning (ML) development environment is crucial for seamless experimentation, training plus testing, and later deployment.

Miniconda, a lightweight distribution of the Conda package manager, offers a minimalistic yet industry-grade powerful approach to constructing such ML platforms.

This article will guide you through the process of setting up a basic machine-learning environment using Miniconda and Python.

But first…

Why Miniconda, and not Anaconda?

Miniconda is a distribution of Conda, a cross-platform package manager that simplifies the installation and management of software packages and dependencies. Unlike Anaconda, which comes bundled with a comprehensive collection of pre-installed packages, Miniconda provides a barebones installation, allowing users to select and install only the packages they need.

I have used Anaconda in the past when I began learning ML. But I am not a fan of it at all. It slows down the computer significantly and installs many unwanted packages and stuff. Even installing or upgrading something using Anaconda looks easy at first because of its GUI nature, but almost always hits a roadblock due to multiple dependencies issues.

In the following, I will guide the Miniconda-based installation of Jupyter Lab, Python, and relevant ML and Python…