Skip to content

Git and GitHub Setup and Essentials

This tutorial shows you how to set up and use Git and GitHub for effective version control and collaboration in machine learning projects. You'll learn essential commands, authentication setup, and workflows that form the foundation of professional ML development.

Estimated time: 45 minutes

Why This Matters

Problem statement: Managing code changes, collaborating with teammates, and maintaining project history becomes impossible without proper version control as ML projects grow in complexity.

Practical benefits: Git and GitHub enable you to track experiments, collaborate safely with teammates, maintain reproducible ML workflows, and recover from mistakes without losing work. These skills are essential for any ML role where you'll work on shared codebases and need to manage model iterations.

Professional context: Every ML team uses Git for code management, GitHub for collaboration, and version control for tracking model experiments and dataset changes. Mastering these tools early accelerates your integration into professional ML workflows.

Git Workflow

Prerequisites & Learning Objectives

Required knowledge: - Basic command line navigation - Understanding of files and folders - GitHub account (create at github.com)

Required tools: - Git installed on your system - Command line access (Terminal/Command Prompt) - Text editor or IDE

Learning outcomes: - Configure Git with your identity and GitHub authentication - Create, stage, commit, and push changes to repositories - Clone existing projects and collaborate through pull/push workflows - Troubleshoot common Git authentication and workflow issues

High-level approach: You'll configure Git locally, set up secure GitHub authentication, practice core workflow commands, and connect local repositories to GitHub for collaboration.

Step-by-Step Instructions

Step 1: Configure Your Git Identity

Git needs to know who you are to properly track changes and contributions.

git config --global user.name "Your Full Name"
git config --global user.email "your-email@example.com"

Why this matters: These details appear in your commit history and help teammates identify who made changes. Use the same email associated with your GitHub account for consistency.

Expected output: No output means success. Verify with: 

git config --global --list

Step 2: Set Up GitHub Authentication

GitHub requires secure authentication through Personal Access Tokens (PATs) instead of passwords.

Create your PAT: - Log in to GitHub - Navigate to Settings → Developer settings → Personal access tokens → Tokens (classic) - Click Generate new token (classic) - Select scopes: repo, workflow, write:packages - Set expiration (no expiration, or 90 days) - Copy the token immediately - See short video tutorial

Store your token securely: Save it in a password manager or secure note - you won't see it again.

Expected outcome: You have a PAT that starts with ghp_ and is 40+ characters long.

Step 3: Create Your Repository on GitHub First

The most streamlined approach is to create your repository directly on GitHub, then clone it to your local machine. This eliminates potential conflicts and simplifies the authentication process.

Create a new repository on GitHub: - Navigate to github.com and click the green "New" button - Name your repository (e.g., my-ml-project) - Keep it public for learning purposes - Important: Leave "Initialize this repository with a README" unchecked - Click "Create repository"

Why this approach works better: Creating an empty repository first gives you a clean foundation and GitHub generates the exact clone URL you need, eliminating configuration errors.

Step 4: Clone Your Repository Locally

Since you already have your Personal Access Token, you can clone the repository using the embedded authentication method:

git clone https://username:token@github.com/username/repository-name.git

**Make sure to replace username, token and repository-name with your actual data.

Example: 

git clone https://your-username:ghp_your_token_here@github.com/your-username/my-ml-project.git
cd my-ml-project

Why embed the token: This method stores your credentials temporarily and eliminates repeated authentication prompts during your session.

Expected output: 

Cloning into 'my-ml-project'...
remote: Enumerating objects: 3, done.
remote: Total 3 (delta 0), reused 0 (delta 0), pack-reused 0
Receiving objects: 100% (3/3), done.

Step 5: Master the Essential Git Workflow

Now that your repository is connected, follow the fundamental pattern: create/modify → stage → commit → push.

Create your first file: 

echo "# My ML Project" > README.md # this creates a file with heading "My ML Project"
echo "This repository contains my machine learning experiments." >> README.md

Check what's changed: 

git status

Stage your changes: 

git add README.md

Commit with a descriptive message: 

git commit -m "Add project README with initial description"

Push to GitHub: 

git push origin main

Why this workflow matters: This four-step process (modify → add → commit → push) forms the essential of version control. Each commit creates a checkpoint you can return to, and pushing synchronizes your work with GitHub.

Expected output after pushing: 

Enumerating objects: 3, done.
Counting objects: 100% (3/3), done.
Writing objects: 100% (3/3), 245 bytes | 245.00 KiB/s, done.
Total 3 (delta 0), reused 0 (delta 0)
To https://github.com/username/my-ml-project.git
 * [new branch]      main -> main

Pro tip: Always use present-tense commit messages ("Add feature" not "Added feature") to maintain consistency with Git's own messaging style.

Quick reference for daily use

Command Purpose When to Use
git status Check what changed Before staging files
git add . Stage all changes When ready to commit everything
git commit -m "message" Save changes locally After staging files
git push Upload to GitHub After committing changes
git pull Download latest changes Before starting new work

Summary & Next Steps

Key accomplishments: You've configured Git with your identity, set up secure GitHub authentication, mastered the core workflow (stage → commit → push), and can now collaborate on repositories professionally.

Best practices for ML development: - Commit frequently with descriptive messages to track experiment iterations - Pull before pushing to avoid conflicts when collaborating
- Use .gitignore to exclude large model files and sensitive data - Branch for experiments to keep main branch stable during model development

External resources for deeper learning: