Package ai.djl.training.listener

Interface TrainingListener.Defaults

Enclosing interface:

TrainingListener

public static interface TrainingListener.Defaults

Contains default TrainingListener sets.

Method Summary

Static Methods

Modifier and Type	Method	Description
static TrainingListener[]	basic()	A basic TrainingListener set with the minimal recommended functionality.
static TrainingListener[]	logging()	A default TrainingListener set including batch output logging.
static TrainingListener[]	logging(int frequency)	A default TrainingListener set including batch output logging.
static TrainingListener[]	logging(String outputDir)	A default TrainingListener set including batch output logging and output directory.

Method Details

basic

static TrainingListener[] basic()

A basic TrainingListener set with the minimal recommended functionality.

This contains:

• EpochTrainingListener
• EvaluatorTrainingListener
• DivergenceCheckTrainingListener

Returns:

the new set of listeners

logging

static TrainingListener[] logging()

A default TrainingListener set including batch output logging.

This contains:

• Everything from basic()
• LoggingTrainingListener

Returns:

the new set of listeners

logging

static TrainingListener[] logging(int frequency)

A default TrainingListener set including batch output logging.

This has the same listeners as logging(), but reduces the logging frequency.

Parameters:

frequency - the frequency of epoch to print out

Returns:

the new set of listeners

logging

static TrainingListener[] logging(String outputDir)

A default TrainingListener set including batch output logging and output directory.

This contains:

• Everything from logging()
• MemoryTrainingListener
• TimeMeasureTrainingListener

Parameters:

outputDir - the output directory to store created log files. Can't be null

Returns:

the new set of listeners