Server

class Server(*, client_manager: ClientManager, strategy: Strategy | None = None)

기반 클래스: object

Flower 서버.

메소드

client_manager()	Return ClientManager.
disconnect_all_clients(timeout)	Send shutdown signal to all clients.
evaluate_round(server_round, timeout)	Validate current global model on a number of clients.
fit(num_rounds, timeout)	Run federated averaging for a number of rounds.
fit_round(server_round, timeout)	Perform a single round of federated averaging.
set_max_workers(max_workers)	Set the max_workers used by ThreadPoolExecutor.
set_strategy(strategy)	Replace server strategy.

client_manager() → ClientManager

Return ClientManager.

disconnect_all_clients(timeout: float | None) → None

Send shutdown signal to all clients.

evaluate_round(server_round: int, timeout: float | None) → tuple[float | None, dict[str, bool | bytes | float | int | str], tuple[list[tuple[ClientProxy, EvaluateRes]], list[tuple[ClientProxy, EvaluateRes] | BaseException]]] | None

Validate current global model on a number of clients.

fit(num_rounds: int, timeout: float | None) → tuple[History, float]

Run federated averaging for a number of rounds.

fit_round(server_round: int, timeout: float | None) → tuple[Parameters | None, dict[str, bool | bytes | float | int | str], tuple[list[tuple[ClientProxy, FitRes]], list[tuple[ClientProxy, FitRes] | BaseException]]] | None

Perform a single round of federated averaging.

set_max_workers(max_workers: int | None) → None

Set the max_workers used by ThreadPoolExecutor.

set_strategy(strategy: Strategy) → None

Replace server strategy.