FedBRB: An Effective Solution to the Small-to-Large Scenario in Device-Heterogeneity Federated Learning

no code implementations • 27 Feb 2024 • Ziyue Xu, Mingfeng Xu, Tianchi Liao, Zibin Zheng, Chuan Chen

FedBRB can uses small local models to train all blocks of the large global model, and broadcasts the trained parameters to the entire space for faster information interaction.

Federated Learning