Introduction

Welcome to Hugging Face 101! This course will teach you the essentials of the Hugging Face ecosystem — home to over 2 billion models and 500,000 dataset.

You’ll learn three ways to work with Hugging Face: through the web interface (no coding required), via APIs (for quick integration), or with Python libraries (for full control). Pick the approach that fits your project, or mix and match as needed!

What You’ll Learn

By the end of this course, you’ll know how to navigate the Hugging Face Hub to find and use the models, datasets, and spaces you need for your projects. You’ll learn to integrate AI capabilities into your applications using APIs for text generation, image creation, and more, without needing to manage complex infrastructure.

The course will teach you how to share your work with the community by uploading your own models, datasets, and interactive demos called Spaces. You’ll discover how to build these Spaces to showcase your AI projects and make them accessible to others.

Additionally, you’ll learn evaluation techniques to test whether AI models work effectively for your specific use case.

Finally, you’ll explore advanced features like the Model Context Protocol (MCP), which enables AI models to connect to external tools and data sources, expanding their capabilities beyond traditional text processing.

Course Structure

This course is divided into 8 chapters, each with a focus on a different aspect of Hugging Face.

1. Introduction
2. Introduction to Hugging Face - Learn about the Hugging Face ecosystem and how to get started
3. Inference Providers - Learn about the different ways to use Hugging Face models
4. Sharing Models - Learn how to share your own models with the community
5. Sharing Datasets - Learn how to share your own datasets with the community
6. Building Spaces - Learn how to build interactive demos with Hugging Face Spaces
7. Building with Gradio - Learn how to build interactive demos with Gradio
8. Evaluating Models - Learn how to evaluate AI models

Why Learn Hugging Face?

Hugging Face has become the go-to platform for AI development. Think of it as the GitHub for AI - where the community shares models, datasets, and applications. Recent innovations like the Model Context Protocol (MCP) are making AI models even more powerful by connecting them to external tools and data.

Whether you’re new to AI, a developer adding AI features to your app, or a researcher sharing your work, Hugging Face gives you the tools and community to succeed.

Choose Your Learning Path

This course works for everyone, from complete beginners to experienced developers. Pick the path that matches your style:

Start with Web Interfaces

Perfect if you’re new to AI or prefer clicking over coding

This path focuses on exploring models through the Hub’s web interface, where you can try AI models directly in your browser without writing any code. You’ll learn to create Spaces (interactive demos) using visual tools and templates.

Deep Dive with Python and JavaScript

Best if you want full control and customization

This comprehensive path covers installing Hugging Face libraries locally, fine-tuning models for your specific needs, and building advanced applications with Python. You’ll gain complete control over the AI pipeline, from data preprocessing to model deployment and monitoring.

Focus on: All chapters, hands-on coding

What You Need to Get Started

Everyone needs:

• A Hugging Face account (free)
• An internet browser
• Curiosity about AI!

For API and Code paths, also helpful:

• Basic familiarity with web APIs (we’ll explain as we go)
• Some Python experience (but beginners can follow along)

Using Google Colab (Recommended for Beginners)

For Code-First and API-First paths. UI-First learners can skip to Chapter 1.

Google Colab is the easiest way to get started - no installation required! Just open a notebook in your browser and start coding.

New to Colab? Check out their introduction first. Colab gives you free access to GPUs and TPUs for AI experiments.

Ready? Create a new notebook and let’s install the libraries:

Install the Hugging Face libraries with pip. In Colab, use ! before pip commands:

# Core libraries for this course
!pip install huggingface_hub transformers datasets gradio

# Optional: MCP client and tooling support
!pip install -U "huggingface_hub[mcp]"

Test that everything installed correctly:

import huggingface_hub
import transformers
import datasets
import gradio

print("✅ Ready to go!")

What Each Library Does

The huggingface_hub library serves as your primary connection to the Hugging Face Hub, handling authentication, model downloads, and uploads. It manages the communication between your local environment and the vast repository of AI models and datasets. When you install the optional mcp extra, it also provides support for the Model Context Protocol, enabling advanced AI agent capabilities.

The transformers library provides access to thousands of pre-trained models spanning text, images, audio, and multimodal tasks. It offers a unified API for loading, using, and fine-tuning state-of-the-art models like BERT, GPT, and CLIP, making it easy to switch between different architectures for your specific needs.

The datasets library simplifies working with machine learning datasets. It provides efficient loading and processing of popular datasets, along with tools for preparing your own data. The library handles memory management automatically, making it possible to work with datasets that are too large to fit in memory.

Finally, gradio enables you to build interactive web demos with minimal code. You can create user-friendly interfaces for your AI models, allowing others to test and interact with your work directly through a web browser.

To use the Model Context Protocol (MCP) client and tiny-agents, install the mcp extra: pip install -U "huggingface_hub[mcp]".

You’re now ready for all the projects in this course!

Local Python Setup

For experienced developers who want to work locally instead of Colab.

If you prefer local development, first install Python from python.org or follow this detailed guide.

Verify Python is installed:

python --version

We recommend using virtual environments to avoid package conflicts. Think of them as isolated Python setups for different projects.

Create a project directory and virtual environment:

# Create and enter your project directory
mkdir ~/huggingface-course
cd ~/huggingface-course

# Create virtual environment
python -m venv .env

# Activate it (Mac/Linux)
source .env/bin/activate

# On Windows, use:
# .env\Scripts\activate

Create a directory for your project and a virtual environment:

# Create and enter your project directory
mkdir ~/huggingface-course
cd ~/huggingface-course

With uv you can set up a project with a virtual environment and install the libraries:

uv init

We need to create a virtual environment:

uv venv

Then activate the virtual environment:

source .venv/bin/activate

Now we can add the libraries to our project using uv add which will add the libraries to the pyproject.toml file and install them in the virtual environment.

uv add install transformers datasets gradio "huggingface_hub[mcp]"

uv will now manage

When activated, your terminal prompt will change to show (.env). To deactivate later, just run deactivate.

Install the Libraries

Now install the same Hugging Face libraries:

pip install huggingface_hub transformers datasets gradio
pip install -U "huggingface_hub[mcp]"  # Optional MCP support

Test the installation:

python -c "import huggingface_hub, transformers, datasets, gradio; print('✅ Ready to go!')"

Get Your Access Token

To upload models, create private repos, and access gated models, you’ll need to authenticate:

from huggingface_hub import login
login()

This will ask for your Hugging Face token. Get yours from huggingface.co/settings/tokens.

What’s Next?

You’re ready to start! Here’s where to go based on your chosen path.