These are the exercises for the hands-on sessions at the Generative Modeling Spring School 2026.
The material consists of two Jupyter notebooks:
1-VAE.ipynb: Variational Autoencoders, used in the Monday session2-FM.ipynb: Flow Matching, used in the Tuesday session
The notebooks run out of the box in Google Colab, but they can also be run locally with minimal setup.
Prerequisites
- The background material presented in the lectures and the book.
- Programming experience in Python using PyTorch