Package org.deeplearning4j.iterator

Class CnnSentenceDataSetIterator.Builder

java.lang.Object

org.deeplearning4j.iterator.CnnSentenceDataSetIterator.Builder

Enclosing class:

CnnSentenceDataSetIterator

public static class CnnSentenceDataSetIterator.Builder
extends Object

Constructor Summary

Constructors

Constructor	Description
Builder()	Deprecated. Due to old default, that will be changed in the future.
Builder(@NonNull CnnSentenceDataSetIterator.Format format)

Method Summary

Modifier and Type	Method	Description
CnnSentenceDataSetIterator	build()
CnnSentenceDataSetIterator.Builder	dataSetPreProcessor(org.nd4j.linalg.dataset.api.DataSetPreProcessor dataSetPreProcessor)	Optional DataSetPreProcessor
CnnSentenceDataSetIterator.Builder	maxSentenceLength(int maxSentenceLength)	Maximum sentence/document length.
CnnSentenceDataSetIterator.Builder	minibatchSize(int minibatchSize)	Minibatch size to use for the DataSetIterator
CnnSentenceDataSetIterator.Builder	sentenceProvider(LabeledSentenceProvider labeledSentenceProvider)	Specify how the (labelled) sentences / documents should be provided
CnnSentenceDataSetIterator.Builder	sentenceProvider(LabelAwareDocumentIterator iterator, @NonNull List<String> labels)	Specify how the (labelled) sentences / documents should be provided
CnnSentenceDataSetIterator.Builder	sentenceProvider(LabelAwareIterator iterator, @NonNull List<String> labels)	Specify how the (labelled) sentences / documents should be provided
CnnSentenceDataSetIterator.Builder	sentenceProvider(LabelAwareSentenceIterator iterator, @NonNull List<String> labels)	Specify how the (labelled) sentences / documents should be provided
CnnSentenceDataSetIterator.Builder	sentencesAlongHeight(boolean sentencesAlongHeight)	If true (default): output features data with shape [minibatchSize, 1, maxSentenceLength, wordVectorSize] If false: output features with shape [minibatchSize, 1, wordVectorSize, maxSentenceLength]
CnnSentenceDataSetIterator.Builder	tokenizerFactory(TokenizerFactory tokenizerFactory)	The TokenizerFactory that should be used.
CnnSentenceDataSetIterator.Builder	unknownWordHandling(CnnSentenceDataSetIterator.UnknownWordHandling unknownWordHandling)	Specify how unknown words (those that don't have a word vector in the provided WordVectors instance) should be handled.
CnnSentenceDataSetIterator.Builder	useNormalizedWordVectors(boolean useNormalizedWordVectors)	Whether normalized word vectors should be used.
CnnSentenceDataSetIterator.Builder	wordVectors(WordVectors wordVectors)	Provide the WordVectors instance that should be used for training

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

@Deprecated
public Builder()

Deprecated.

Due to old default, that will be changed in the future. Use Builder(Format) to specify the CnnSentenceDataSetIterator.Format of the activations

Builder

public Builder(@NonNull
               @NonNull CnnSentenceDataSetIterator.Format format)

Parameters:

format - The format to use for the features - i.e., for 1D or 2D CNNs

Method Detail

sentenceProvider

public CnnSentenceDataSetIterator.Builder sentenceProvider(LabeledSentenceProvider labeledSentenceProvider)

Specify how the (labelled) sentences / documents should be provided

sentenceProvider

public CnnSentenceDataSetIterator.Builder sentenceProvider(LabelAwareIterator iterator,
                                                           @NonNull
                                                           @NonNull List<String> labels)

Specify how the (labelled) sentences / documents should be provided

sentenceProvider

public CnnSentenceDataSetIterator.Builder sentenceProvider(LabelAwareDocumentIterator iterator,
                                                           @NonNull
                                                           @NonNull List<String> labels)

Specify how the (labelled) sentences / documents should be provided

sentenceProvider

public CnnSentenceDataSetIterator.Builder sentenceProvider(LabelAwareSentenceIterator iterator,
                                                           @NonNull
                                                           @NonNull List<String> labels)

Specify how the (labelled) sentences / documents should be provided

wordVectors

public CnnSentenceDataSetIterator.Builder wordVectors(WordVectors wordVectors)

Provide the WordVectors instance that should be used for training

tokenizerFactory

public CnnSentenceDataSetIterator.Builder tokenizerFactory(TokenizerFactory tokenizerFactory)

The TokenizerFactory that should be used. Defaults to DefaultTokenizerFactory

unknownWordHandling

public CnnSentenceDataSetIterator.Builder unknownWordHandling(CnnSentenceDataSetIterator.UnknownWordHandling unknownWordHandling)

Specify how unknown words (those that don't have a word vector in the provided WordVectors instance) should be handled. Default: remove/ignore unknown words.

minibatchSize

public CnnSentenceDataSetIterator.Builder minibatchSize(int minibatchSize)

Minibatch size to use for the DataSetIterator

useNormalizedWordVectors

public CnnSentenceDataSetIterator.Builder useNormalizedWordVectors(boolean useNormalizedWordVectors)

Whether normalized word vectors should be used. Default: true

maxSentenceLength

public CnnSentenceDataSetIterator.Builder maxSentenceLength(int maxSentenceLength)

Maximum sentence/document length. If sentences exceed this, they will be truncated to this length by taking the first 'maxSentenceLength' known words.

sentencesAlongHeight

public CnnSentenceDataSetIterator.Builder sentencesAlongHeight(boolean sentencesAlongHeight)

If true (default): output features data with shape [minibatchSize, 1, maxSentenceLength, wordVectorSize]
If false: output features with shape [minibatchSize, 1, wordVectorSize, maxSentenceLength]

dataSetPreProcessor

public CnnSentenceDataSetIterator.Builder dataSetPreProcessor(org.nd4j.linalg.dataset.api.DataSetPreProcessor dataSetPreProcessor)

Optional DataSetPreProcessor

build

public CnnSentenceDataSetIterator build()