This is work has been accepted to the IEEE Transactions on Network and Service Management, Special Issue on "Research Advances Towards Effective and Sustainable Next Generation Networks".
This repository contains a discrete-event simulator capable of simulating Gossip Learning (GL) with various network topologies and stopping criteria.
All the experimental material related to the paper and the assets are available on Zenodo.
It contains the code of the Python simulator package, utility scripts to run a simulation, and numbered Jupyter Notebooks step-by-step perform an end-to-end simulation (e.g., from dataset preparation to evaluation).
It contains the code of a Flower app used as the baseline for Federated Learning experiments.
Distributed under the GPL v3 license. See LICENSE for more information.