armlet.utils.log module#

class armlet.utils.log.ArmletLog(json_log_dir: str | None = None, **kwargs)#

Bases: Log

close() → None#

Close the logger.

end_fit(round: int, client_id: int, model: Module, loss: float, **kwargs) → None#

This method is called when the client ends the local training process.

Parameters:

  • round (int) – The round number.

  • client_id (int) – The client ID.

  • model (Module) – The local model after the local training.

  • loss (float) – The average loss incurred by the local model during training.

  • **kwargs (dict) – Additional keyword arguments.