ai.djl.training

Class EasyTrain

java.lang.Object

ai.djl.training.EasyTrain

public final class EasyTrain
extends java.lang.Object

Helper for easy training of a whole model, a trainining batch, or a validation batch.

Method Summary

Modifier and Type	Method and Description
static void	fit(Trainer trainer, int numEpoch, Dataset trainingDataset, Dataset validateDataset) Runs a basic epoch training experience with a given trainer.
static void	trainBatch(Trainer trainer, Batch batch) Trains the model with one iteration of the given Batch of data.
static void	validateBatch(Trainer trainer, Batch batch) Validates the given batch of data.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

fit

public static void fit(Trainer trainer,
                       int numEpoch,
                       Dataset trainingDataset,
                       Dataset validateDataset)
                throws java.io.IOException,
                       TranslateException

Runs a basic epoch training experience with a given trainer.

Parameters:

trainer - the trainer to train for

numEpoch - the number of epochs to train

trainingDataset - the dataset to train on

validateDataset - the dataset to validate against. Can be null for no validation

Throws:

java.io.IOException - for various exceptions depending on the dataset

TranslateException - if there is an error while processing input

trainBatch

public static void trainBatch(Trainer trainer,
                              Batch batch)

Trains the model with one iteration of the given Batch of data.

Parameters:

trainer - the trainer to validate the batch with

batch - a Batch that contains data, and its respective labels

Throws:

java.lang.IllegalArgumentException - if the batch engine does not match the trainer engine

validateBatch

public static void validateBatch(Trainer trainer,
                                 Batch batch)

Validates the given batch of data.

During validation, the evaluators and losses are computed, but gradients aren't computed, and parameters aren't updated.

Parameters:

trainer - the trainer to validate the batch with

batch - a Batch of data

Throws:

java.lang.IllegalArgumentException - if the batch engine does not match the trainer engine