MLFlow Mod

Contents

MLFlow Mod#

rizemind.logging.mlflow.mod.mlflow_mod(msg: Message, ctx: Context, call_next: Callable[[Message, Context], Message]) → Message[source]#

Logs metrics on an incoming TRAIN message to an Mlflow server.

The mlflow_mod relies on the TRAIN_METRIC_HISTORY_KEY as a standardized metric type and reads the content of this metric for logging. In addition to the metrics available in TRAIN_METRIC_HISTORY_KEY, mlflow_mod automatically logs training_time and epochs.

Parameters:

  • msg – The incoming message from the ServerApp to the ClientApp.

  • ctx – Context of the run.

  • call_next – The next callable in the chain to process the message.

Returns:

The response message sent from the ClientApp to the ServerApp.

Contents