intro_dgm

Why we included this project

This repository pairs each chapter of Jakub Tomczak's Springer book on deep generative modeling with a short Jupyter notebook, and together they span nearly every major model family: mixture models, autoregressive models, flows, VAEs, diffusion and score-based models, EBMs, GANs, neural compression, and a small decoder-only transformer. The notebooks are deliberately minimal and run in minutes on a laptop, so a newcomer can see working PyTorch code for each technique before building their own. Students, self-learners, and instructors all get value here, with ready-made assignment examples and lecture figures for the classroom. It is teaching material rather than production tooling, but it is a practical reference for a small team getting up to speed on modern generative AI.