Tutorials on MACE training and architecture

We have built a series of tutorials to help you get started with MACE. These tutorials are designed to help you understand the basics of MACE, how to train a model, and how to use it in your own projects. The tutorials were made by Ioan Magdău, Ilyes Batatia and Will Baldwin. All tutorials are available on Google Colab, so you can run them in your browser without any setup. If you want want to run the tutorials locally, you can download the notebooks from Google Colab.

Tutorial 1: Introduction to MACE training and evaluation

In this tutorial, we will introduce you to the basics of MACE training and evaluation. We cover the construction of a dataset, the basic hyperparameters of a MACE model, and how to train and evaluate a model.

Link to tutorial in colab: https://colab.research.google.com/drive/1ZrTuTvavXiCxTFyjBV4GqlARxgFwYAtX

Tutorial 2: MACE active learning and fine-tuning

In this tutorial, we will show you how to use MACE for active learning and fine-tuning.

Link to tutorial in colab: https://colab.research.google.com/drive/1oCSVfMhWrqHTeHbKgUSQN9hTKxLzoNyb

Tutorial 3: MACE theory and code (advanced)

In this tutorial, we will dive into the theory and code of MACE. Each section of the code is explained in detail, and we reference the corresponding equations in the manuscript.

Link to tutorial in colab: https://colab.research.google.com/drive/1AlfjQETV_jZ0JQnV5M3FGwAM2SGCl2aU