scaffold.torch.lightning.loggers

Attributes

logger

Classes

HyperoptLightningLogger

Lightning wrapper around hyperopt logging.

Module Contents

class scaffold.torch.lightning.loggers.HyperoptLightningLogger

Bases: pytorch_lightning.loggers.logger.Logger

Lightning wrapper around hyperopt logging.

finalize(status: str) → None

Finalisation of experiment run.

Parameters:

status (str) – The experiment outcome (success, failed, etc.).

log_hyperparams(params: Dict[str, Any] | argparse.Namespace) → None

Logging of hyperparameters for experiment.

Parameters:

params (argparse.Namespace) – The parameters to be logged.

log_metrics(metrics: Dict[str, Any], step: int) → None

Logging of metrics.

Parameters:

• metrics (Dict) – Dictionary of metric names and values to be logged

• step (int) – The step counter of optimisation iterations

property name: str

The experiment name.

property version: str

The experiment version.

scaffold.torch.lightning.loggers.logger