Welcome to the UvA Deep Learning Tutorials!

Abstraction: University of Amsterdam Jupyter notebook deep learning course covering PyTorch and JAX

Key points:

• Created by Phillip Lippe for UvA DL1 (Fall 2025) and DL2 (Spring 2026); integrated as official PyTorch Lightning tutorials
• 17+ tutorial notebooks covering activation functions, optimization/initialization, ResNet/DenseNet, Transformers, GNNs, autoencoders, normalizing flows, vision transformers, meta-learning, and SimCLR
• Notebooks available in both PyTorch and JAX+Flax versions; runnable locally on CPU, Google Colab, or Snellius cluster
• DL2 track adds geometric deep learning, Bayesian NNs with Pyro, dynamical systems/Neural ODEs, and causal representation learning
• Distributed/scaled training tutorials cover data parallelism, pipeline parallelism, tensor parallelism, and 3D parallelism in JAX
• Pretrained models auto-downloaded; total disk requirement under 1GB

Connections: Pytorch · Deep Learning · Transformers · Graph Neural Networks · Normalizing Flows

Source: https://uvadlc-notebooks.readthedocs.io/en/latest/index.html