Package ai.djl.training.listener
Contains classes providing functionality during training through
TrainingListener
.
It contains a main interface TrainingListener
and various
listeners that extend it.
-
Interface Summary Interface Description TrainingListener TrainingListener
offers an interface that performs some actions when certain events have occurred in theTrainer
.TrainingListener.Defaults Contains defaultTrainingListener
sets. -
Class Summary Class Description DivergenceCheckTrainingListener TrainingListener
that gives early warning if your training has failed by divergence.EarlyStoppingListener Listener that allows the training to be stopped early if the validation loss is not improving, or if time has expired.EarlyStoppingListener.Builder A builder for aEarlyStoppingListener
.EpochTrainingListener EpochTrainingListener
that tracks epochs.EvaluatorTrainingListener TrainingListener
that records evaluator results.LoggingTrainingListener TrainingListener
that outputs the progress of training each batch and epoch into logs.MemoryTrainingListener TrainingListener
that collects the memory usage information.SaveModelTrainingListener ATrainingListener
that saves a model and can save checkpoints.TimeMeasureTrainingListener TrainingListener
that outputs the training time metrics after training is done.TrainingListener.BatchData A class to pass data from the batch into the training listeners.TrainingListenerAdapter Base implementation of the training listener that does nothing. -
Exception Summary Exception Description EarlyStoppingListener.EarlyStoppedException Thrown when training is stopped early, the message will contain the reason why it is stopped early.