FedProx

Implementation of the [FedProx18] algorithm.

References

[FedProx18]

Tian Li, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, and Virginia Smith. Federated Optimization in Heterogeneous Networks. Adaptive & Multitask Learning Workshop. URL: https://openreview.net/pdf?id=SkgwE5Ss3N

Classes included in fluke.algorithms.fedprox

FedProxClient
FedProx

Classes

class fluke.algorithms.fedprox.FedProxClient(index: int, train_set: FastDataLoader, test_set: FastDataLoader, optimizer_cfg: OptimizerConfigurator, loss_fn: Module, local_epochs: int, mu: float, fine_tuning_epochs: int = 0, clipping: float = 0, **kwargs)

Bases: Client

class fluke.algorithms.fedprox.FedProx(n_clients: int, data_splitter: DataSplitter, hyper_params: DDict | dict[str, Any], **kwargs)

Bases: CentralizedFL