Result¶
- class Result(arrays: ~flwr.common.record.arrayrecord.ArrayRecord = <factory>, train_metrics_clientapp: dict[int, ~flwr.common.record.metricrecord.MetricRecord] = <factory>, evaluate_metrics_clientapp: dict[int, ~flwr.common.record.metricrecord.MetricRecord] = <factory>, evaluate_metrics_serverapp: dict[int, ~flwr.common.record.metricrecord.MetricRecord] = <factory>)[source]¶
Bases:
object
Data class carrying records generated during the execution of a strategy.
This class encapsulates the results of a federated learning strategy execution, including the final global model parameters and metrics collected throughout the federated training and evaluation (both federated and centralized) stages.
- arrays¶
The final global model parameters. Contains the aggregated model weights/parameters that resulted from the federated learning process.
- Type:
- train_metrics_clientapp¶
Training metrics collected from ClientApps, indexed by round number. Contains aggregated metrics (e.g., loss, accuracy) from the training phase of each federated learning round.
- Type:
dict[int, MetricRecord]
- evaluate_metrics_clientapp¶
Evaluation metrics collected from ClientApps, indexed by round number. Contains aggregated metrics (e.g. validation loss) from the evaluation phase where ClientApps evaluate the global model on their local validation/test data.
- Type:
dict[int, MetricRecord]
- evaluate_metrics_serverapp¶
Evaluation metrics generated at the ServerApp, indexed by round number. Contains metrics from centralized evaluation performed by the ServerApp (e.g., when the server evaluates the global model on a held-out dataset).
- Type:
dict[int, MetricRecord]
Methods
Attributes