gossipy documentation#
Under construction!
This documentation is currently under construction.
GossiPy is a python-based open-source framework for performing Gossip Learning (GL) and decentralized federated learning (DFL). GossiPy is designed by researchers for researchers and hence the overall architecture aims at being easy to extend but also easy to use off-the-shelf. As far as we know is the first python module for simulating gossip learning and decentralized federated learning. GossiPy has been developed following the General Responsibility Assignment Software Principles (GRASP) with a particular focus on Low Coupling and High Cohesion. This python module has been written following PEP 8 – Style Guide for Python Code.