Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/ageron/handson-ml3/llms.txt

Use this file to discover all available pages before exploring further.

Hands-On Machine Learning (3rd edition) is the companion notebook repository for the O’Reilly book by Aurélien Géron. It provides complete, runnable Jupyter notebooks for every chapter — covering the full arc from classical ML with Scikit-Learn through modern deep learning with Keras and TensorFlow 2. Each notebook includes working code, visualizations, and exercise solutions.

Quickstart

Run notebooks in the cloud in one click — no installation needed

Installation

Set up a local environment with Anaconda and Jupyter

Part I: ML Fundamentals

Classification, SVMs, trees, ensembles, and unsupervised learning

Part II: Deep Learning

Neural networks, CNNs, RNNs, transformers, and generative models

What’s inside

The repository contains 19 main notebooks — one per book chapter — plus supplementary tutorials on scientific Python and math prerequisites.

Scikit-Learn

End-to-end ML projects, classification, regression, and clustering

TensorFlow & Keras

Build, train, and tune neural networks with the Keras API

Computer Vision

Image classification and object detection with CNNs

NLP & Transformers

Sequence models, attention mechanisms, and language models

Generative Models

Autoencoders, GANs, and diffusion models

Production ML

Deploy models at scale with Vertex AI and TensorFlow Serving

Run notebooks instantly

Open in Google Colab

The recommended way to run notebooks — no setup required

Open in Kaggle

Run notebooks directly in Kaggle’s free environment

Getting started locally

1

Clone the repository

git clone https://github.com/ageron/handson-ml3.git
cd handson-ml3
2

Create the conda environment

conda env create -f environment.yml
conda activate homl3
3

Register the Jupyter kernel

python -m ipykernel install --user --name=python3
4

Launch Jupyter

jupyter notebook
Open index.ipynb in your browser to navigate all chapters.
If you have a TensorFlow-compatible NVIDIA GPU, see the installation guide for GPU setup instructions including CUDA and cuDNN.

Notebook overview

#ChapterTopics
1The ML LandscapeTypes of ML, model selection, challenges
2End-to-End ML ProjectPipeline, feature engineering, evaluation
3ClassificationBinary/multiclass, precision/recall, ROC
4Training ModelsLinear/polynomial regression, regularization
5Support Vector MachinesLinear SVM, kernel trick, SVR
6Decision TreesCART, pruning, visualization
7Ensemble LearningBagging, boosting, XGBoost
8Dimensionality ReductionPCA, t-SNE, UMAP
9Unsupervised LearningK-Means, DBSCAN, Gaussian mixtures
10Neural Nets with KerasSequential/Functional API, callbacks
11Training Deep NetworksBatch norm, dropout, learning schedules
12Custom TF ModelsCustom layers, training loops, autodiff
13Data Pipelinestf.data, TF Datasets, preprocessing layers
14Computer Vision CNNsResNet, EfficientNet, object detection
15Sequences & Time SeriesRNNs, LSTMs, WaveNet, ARMA
16NLP & AttentionTransformers, BERT, T5, translation
17Autoencoders & GANsVAEs, StyleGAN, DDPM diffusion
18Reinforcement LearningPolicy gradients, DQN, Gymnasium
19Deploy at ScaleVertex AI, TF Serving, distributed training

Build docs developers (and LLMs) love