handson-ml3
A series of Jupyter notebooks that walk you through the fundamentals of Machine Learning and Deep Learning in Python using Scikit-Learn, Keras and TensorFlow 2.
Runnable Jupyter Notebooks from the third edition of Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, including all code examples and exercise solutions from the book.
This repository contains the Jupyter Notebooks from the third edition of the book "Hands-on Machine Learning with Scikit-Learn, Keras and TensorFlow" by Aurelien Geron, published by O'Reilly. The notebooks are the practical, code-alongside-text companion to that book. If you own or are studying from the third edition, this is where you find the example code and the solutions to the exercises.
Machine learning is the field of getting software to learn patterns from data rather than following explicit rules. The book and notebooks cover the fundamentals using three Python tools that are widely used in the field: Scikit-Learn for classic machine learning, and Keras and TensorFlow for deep learning (neural networks). The content is aimed at people who want hands-on experience working through concrete examples in Python.
You do not need to install anything to get started. The notebooks can be opened and run in the browser via Google Colab, which provides free cloud computing. Kaggle, Binder, and Deepnote are listed as alternative online environments that may also work. If you want to just read the notebooks without running code, a static viewer is also available.
For local installation, the README recommends using Anaconda or Miniconda (Python environment tools) along with Git to clone the repository. The steps involve creating a dedicated Python environment from a provided configuration file and then launching Jupyter Notebook. If you have a compatible graphics card you can also set up GPU support for faster training. A detailed installation guide covers edge cases like SSL certificate errors on macOS.
The repository accompanies the third edition specifically. Previous editions have separate repositories linked from the README. Python 3.10 is recommended, though any version 3.7 or newer is expected to work. The project was supported in part by Google Cloud credits and has received contributions from multiple reviewers and developers.
Where it fits
- Work through hands-on machine learning exercises from the book with runnable code, all in a browser via Google Colab at no cost.
- Study neural network implementations in Keras and TensorFlow using the book's example notebooks side by side with the text.
- Check your answers to the book's exercises using the included solution notebooks.
- Set up a local Python environment with optional GPU support for training the book's deep learning models faster.