// the find
ageron/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.
The companion notebooks for Aurélien Géron's 'Hands-On Machine Learning' O'Reilly book, 3rd edition. Covers the full arc from linear models through CNNs, RNNs, transformers, GANs, diffusion models, and RL using scikit-learn, Keras, and TensorFlow 2. This is textbook material — it's meant to be read alongside the book, not used as standalone reference.
The chapter coverage is genuinely thorough — 19 core notebooks plus extras for autodiff, gradient descent internals, and standalone math primers (linear algebra, calculus). Colab support means zero local setup friction, which matters a lot for beginners. The notebooks are tightly coupled to a well-edited book, so the explanations actually connect to the code rather than being afterthoughts. Docker setup is included and maintained, which saves the usual conda environment archaeology.
It's a book companion first. Without the book, several notebooks are thin on explanation and the exercises don't stand alone well. TensorFlow/Keras as the primary deep learning stack is an increasingly awkward choice in 2025 — PyTorch has won the research half of that war, and the notebooks don't acknowledge it. The transformer and attention coverage (chapter 16) predates the LLM era by enough that it feels dated for anyone who came here hoping to understand modern language models. No GPU memory management guidance beyond 'install CUDA' — painful for anyone with a non-Nvidia setup.