org.deeplearning4j.models.word2vec.wordstore

Class VocabConstructor.Builder<T extends SequenceElement>

java.lang.Object

org.deeplearning4j.models.word2vec.wordstore.VocabConstructor.Builder<T>

Enclosing class:

VocabConstructor<T extends SequenceElement>

public static class VocabConstructor.Builder<T extends SequenceElement>
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
Builder()

Method Summary

Modifier and Type	Method and Description
VocabConstructor.Builder<T>	addSource(SequenceIterator<T> iterator, int minElementFrequency) Adds SequenceIterator for vocabulary construction.
VocabConstructor.Builder<T>	allowParallelTokenization(boolean reallyAllow)
VocabConstructor<T>	build()
VocabConstructor.Builder<T>	enableScavenger(boolean reallyEnable)
VocabConstructor.Builder<T>	fetchLabels(boolean reallyFetch) Sets, if labels should be fetched, during vocab building
VocabConstructor.Builder<T>	setEntriesLimit(int limit) This method sets the limit to resulting vocabulary size.
VocabConstructor.Builder<T>	setIndex(InvertedIndex<T> index)
VocabConstructor.Builder<T>	setStopWords(java.util.Collection<java.lang.String> stopWords)
VocabConstructor.Builder<T>	setTargetVocabCache(VocabCache<T> cache) After temporary internal vocabulary is built, it will be transferred to target VocabCache you pass here
VocabConstructor.Builder<T>	setUnk(T unk)
protected VocabConstructor.Builder<T>	useAdaGrad(boolean useAdaGrad) Defines, if adaptive gradients should be created during vocabulary mastering

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

public Builder()

Method Detail

setEntriesLimit

public VocabConstructor.Builder<T> setEntriesLimit(int limit)

This method sets the limit to resulting vocabulary size. PLEASE NOTE: This method is applicable only if huffman tree is built.

Parameters:

limit -

Returns:

allowParallelTokenization

public VocabConstructor.Builder<T> allowParallelTokenization(boolean reallyAllow)

useAdaGrad

protected VocabConstructor.Builder<T> useAdaGrad(boolean useAdaGrad)

Defines, if adaptive gradients should be created during vocabulary mastering

Parameters:

useAdaGrad -

Returns:

setTargetVocabCache

public VocabConstructor.Builder<T> setTargetVocabCache(@NonNull
                                                       VocabCache<T> cache)

After temporary internal vocabulary is built, it will be transferred to target VocabCache you pass here

Parameters:

cache - target VocabCache

Returns:

addSource

public VocabConstructor.Builder<T> addSource(@NonNull
                                             SequenceIterator<T> iterator,
                                             int minElementFrequency)

Adds SequenceIterator for vocabulary construction. Please note, you can add as many sources, as you wish.

Parameters:

iterator - SequenceIterator to build vocabulary from

minElementFrequency - elements with frequency below this value will be removed from vocabulary

Returns:

setStopWords

public VocabConstructor.Builder<T> setStopWords(@NonNull
                                                java.util.Collection<java.lang.String> stopWords)

fetchLabels

public VocabConstructor.Builder<T> fetchLabels(boolean reallyFetch)

Sets, if labels should be fetched, during vocab building

Parameters:

reallyFetch -

Returns:

setIndex

public VocabConstructor.Builder<T> setIndex(InvertedIndex<T> index)

enableScavenger

public VocabConstructor.Builder<T> enableScavenger(boolean reallyEnable)

setUnk

public VocabConstructor.Builder<T> setUnk(T unk)

build

public VocabConstructor<T> build()