// the find
rasbt/machine-learning-book
Code Repository for Machine Learning with PyTorch and Scikit-Learn
The companion code repository for Sebastian Raschka's 'Machine Learning with PyTorch and Scikit-Learn' book (Packt, 2022). It covers the full arc from classical ML with scikit-learn through CNNs, RNNs, transformers, GANs, GNNs, and reinforcement learning. Aimed at developers and students who own the book and want runnable notebooks alongside the theory.
Each chapter ships both a .ipynb and a converted .py script, so you can run code without Jupyter if you prefer. The progression is disciplined — chapter 11 implements a neural net from scratch before chapter 12 hands you PyTorch, which is the right order. Raschka is meticulous about errata and has maintained the repo actively (last push June 2026, four years after publication). Coverage of GNNs and RL in a single volume is rarer than it should be.
This is book companion code, not a standalone learning resource — the notebooks are nearly useless without the prose and formulas from the paid book. The book was written in 2022 and the PyTorch APIs have moved; some patterns (DataLoader usage, Lightning trainer API) are showing their age. There is no automated test suite or CI, so notebook rot is a real risk as dependencies update. The transformer chapter covers GPT-2 and BERT fine-tuning but predates the LLM era entirely — anyone coming for practical LLM work will hit a dead end at chapter 16.