Nineteen chapters of ML, from perceptrons to GNNs

What it does This repository contains the Jupyter notebooks for Machine Learning with PyTorch and Scikit-Learn, a 770-page Packt textbook by Sebastian Raschka and co-authors. The nineteen chapters span scikit-learn classifiers, data preprocessing, and ensemble methods before moving into PyTorch, CNNs, RNNs, Transformers, GANs, graph neural networks, and reinforcement learning. The authors note that these are code excerpts meant to accompany the book’s formulae and explanatory text.

The interesting bit The repo follows the book’s pedagogical chapter order, implementing a multi-layer neural network from scratch before introducing PyTorch, then advancing through modern architectures like attention mechanisms and graph convolutions. It functions less as a software library and more as a curated learning path rendered in notebook form.

Key highlights

• 19 chapter directories covering classical ML through contemporary deep learning
• From-scratch neural network implementation before any PyTorch abstraction
• Advanced topics included: GANs, Graph Neural Networks, and Transformers
• Community-contributed guide for running examples on Google Colab
• Serbian translation of the print edition available

Caveats

• Authors explicitly warn the notebooks “may not be useful without the formulae and descriptive text” from the book
• Not a standalone course or reusable package; strictly companion material

Verdict A useful reference if you own the book or want a structured index of a comprehensive ML curriculum. Skip it if you need self-contained tutorials or installable libraries.