An open-source framework called MALPOLON has been developed to help researchers build deep learning-based species distribution models (deep-SDMs). Written in Python and based on the PyTorch library, the tool is designed for users with general Python skills, such as ecologists, who want to explore deep learning methods for modelling where species occur.

MALPOLON includes ready-to-use examples, supports both custom and provided datasets, and allows more advanced users to adapt its modular structure for specific experiments. It offers straightforward installation, configuration via simple YAML files, parallel computing, multi-GPU use, and baseline models for comparison. The framework is available on GitHub and PyPi, complete with documentation and examples.

Read the paper here and explore MAMBO's library for more research outputs.

The research behind MALPOLON was partly funded by the European Commission through the MAMBO and GUARDEN sister projects.