Tutorials

This section provides useful tutorial to get started using this library, from JAX and Equinox basics to a first functioning example of a classifier on Entangled-MNIST.

Models and concepts from this library are based on the paper by Badalotti, Baldassi, Mézard, Scardecchia and Zecchina. On the final tutorial, we will be training a 1-layer model on a simple benchmark dataset.

The purpose of this library, however, is more general, as it allows for the quick and intuitive implementation of diverse architectures and learning schemes, loosely based on the principles of the original model.

See the first tutorial to get started.

For issues or questions, reach out to one of the Contributors.