// the find
rasbt/python-machine-learning-book-2nd-edition
The "Python Machine Learning (2nd edition)" book code repository and info resource
The code companion to Sebastian Raschka's 2017 book, covering classical ML with scikit-learn through deep learning with TensorFlow 1.x. This is squarely for beginners and intermediate practitioners who want runnable notebooks alongside the textbook, not a standalone learning resource.
Raschka writes unusually clean didactic code — the notebooks implement algorithms from scratch (ch12 builds a neural net by hand) before reaching for library abstractions, which actually teaches something. The chapter on model evaluation and hyperparameter tuning (ch06) is more thorough than most intro materials. Each chapter ships both a .ipynb and a .py script, so you can run examples without Jupyter. There's a real test suite that executes the notebooks end-to-end via Travis CI, which is rare for book repos.
Frozen in 2017: TensorFlow 1.x session-graph style code throughout, which is dead — TF2/Keras replaced it years ago, and the book explicitly notes a 3rd edition exists. The web deployment chapter (ch09) uses Flask with pickled sklearn models and SQLite, an approach that's fine to understand but bad to copy. No PyTorch coverage at all, which is now the dominant research framework. The notebooks are not self-contained — the README warns they're useless without the book's prose and formulas, so you're paying twice to use this.