Murmura: A Ray-based Framework for Federated and Decentralized Machine Learning

Murmura is a comprehensive Ray-based framework for federated and decentralized machine learning. Built for researchers and developers, it provides tools for distributed machine learning with differential privacy guarantees and flexible network topologies. The framework supports both centralized federated learning and decentralized peer-to-peer learning environments, with features including multiple network topologies, Byzantine-robust aggregation strategies, comprehensive differential privacy support, and intelligent resource management.