ai.djl.training.dataset

Class RandomAccessDataset.BaseBuilder<T extends RandomAccessDataset.BaseBuilder<T>>

java.lang.Object

ai.djl.training.dataset.RandomAccessDataset.BaseBuilder<T>

Direct Known Subclasses:

ArrayDataset.Builder

Enclosing class:

RandomAccessDataset

public abstract static class RandomAccessDataset.BaseBuilder<T extends RandomAccessDataset.BaseBuilder<T>>
extends java.lang.Object

The Builder to construct a RandomAccessDataset.

Field Summary

Fields

Modifier and Type	Field and Description
protected Batchifier	dataBatchifier
protected Device	device
protected Batchifier	labelBatchifier
protected long	limit
protected Pipeline	pipeline
protected int	prefetchNumber
protected Sampler	sampler
protected Pipeline	targetPipeline

Constructor Summary

Constructors

Constructor and Description
BaseBuilder()

Method Summary

Modifier and Type	Method and Description
T	addTargetTransform(Transform transform) Adds the Transform to the target Pipeline to be applied on the labels.
T	addTransform(Transform transform) Adds the Transform to the Pipeline to be applied on the data.
Sampler	getSampler() Gets the Sampler for the dataset.
T	optDataBatchifier(Batchifier dataBatchifier) Sets the Batchifier for the data.
T	optDevice(Device device) Sets the Device.
T	optLabelBatchifier(Batchifier labelBatchifier) Sets the Batchifier for the labels.
T	optLimit(long limit) Sets this dataset's limit.
T	optPipeline(Pipeline pipeline) Sets the Pipeline of Transform to be applied on the data.
T	optPrefetchNumber(int prefetchNumber) Sets the number of batches to prefetch at once.
T	optTargetPipeline(Pipeline targetPipeline) Sets the Pipeline of Transform to be applied on the labels.
protected abstract T	self() Returns this {code Builder} object.
T	setSampling(int batchSize, boolean random) Sets the Sampler with the given batch size.
T	setSampling(int batchSize, boolean random, boolean dropLast) Sets the Sampler with the given batch size.
T	setSampling(Sampler sampler) Sets the Sampler for the dataset.

Methods inherited from class java.lang.Object

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

Field Detail

sampler

protected Sampler sampler

dataBatchifier

protected Batchifier dataBatchifier

labelBatchifier

protected Batchifier labelBatchifier

pipeline

protected Pipeline pipeline

targetPipeline

protected Pipeline targetPipeline

prefetchNumber

protected int prefetchNumber

limit

protected long limit

device

protected Device device

Constructor Detail

BaseBuilder

public BaseBuilder()

Method Detail

getSampler

public Sampler getSampler()

Gets the Sampler for the dataset.

Returns:

the Sampler

setSampling

public T setSampling(int batchSize,
                     boolean random)

Sets the Sampler with the given batch size.

Parameters:

batchSize - the batch size

random - whether the sampling has to be random

Returns:

this BaseBuilder

setSampling

public T setSampling(int batchSize,
                     boolean random,
                     boolean dropLast)

Sets the Sampler with the given batch size.

Parameters:

batchSize - the batch size

random - whether the sampling has to be random

dropLast - whether to drop the last incomplete batch

Returns:

this BaseBuilder

setSampling

public T setSampling(Sampler sampler)

Sets the Sampler for the dataset.

Parameters:

sampler - the Sampler to be set

Returns:

this BaseBuilder

optDataBatchifier

public T optDataBatchifier(Batchifier dataBatchifier)

Sets the Batchifier for the data.

Parameters:

dataBatchifier - the Batchifier to be set

Returns:

this BaseBuilder

optLabelBatchifier

public T optLabelBatchifier(Batchifier labelBatchifier)

Sets the Batchifier for the labels.

Parameters:

labelBatchifier - the Batchifier to be set

Returns:

this BaseBuilder

optPipeline

public T optPipeline(Pipeline pipeline)

Sets the Pipeline of Transform to be applied on the data.

Parameters:

pipeline - the Pipeline of Transform to be applied on the data

Returns:

this BaseBuilder

addTransform

public T addTransform(Transform transform)

Adds the Transform to the Pipeline to be applied on the data.

Parameters:

transform - the Transform to be added

Returns:

this builder

optTargetPipeline

public T optTargetPipeline(Pipeline targetPipeline)

Sets the Pipeline of Transform to be applied on the labels.

Parameters:

targetPipeline - the Pipeline of Transform to be applied on the labels

Returns:

this BaseBuilder

addTargetTransform

public T addTargetTransform(Transform transform)

Adds the Transform to the target Pipeline to be applied on the labels.

Parameters:

transform - the Transform to be added

Returns:

this builder

optPrefetchNumber

public T optPrefetchNumber(int prefetchNumber)

Sets the number of batches to prefetch at once.

Parameters:

prefetchNumber - the number of batches to prefetch at once

Returns:

this BaseBuilder

optDevice

public T optDevice(Device device)

Sets the Device.

Parameters:

device - the device

Returns:

this BaseBuilder

optLimit

public T optLimit(long limit)

Sets this dataset's limit.

The limit is usually used for testing purposes to test only with a subset of the dataset.

Parameters:

limit - the limit of this dataset's records

Returns:

this BaseBuilder

self

protected abstract T self()

Returns this {code Builder} object.

Returns:

this BaseBuilder