org.deeplearning4j.models.word2vec

Class StaticWord2Vec.Builder

java.lang.Object

org.deeplearning4j.models.word2vec.StaticWord2Vec.Builder

Enclosing class:

StaticWord2Vec

public static class StaticWord2Vec.Builder
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
Builder(org.nd4j.linalg.compression.AbstractStorage<java.lang.Integer> storage, VocabCache<VocabWord> vocabCache)

Method Summary

Modifier and Type	Method and Description
StaticWord2Vec	build() This method returns Static Word2Vec implementation, which is suitable for tasks like neural nets feeding.
StaticWord2Vec.Builder	setCachePerDevice(long bytes) This method lets you to define if decompressed values will be cached, to avoid excessive decompressions.

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

public Builder(org.nd4j.linalg.compression.AbstractStorage<java.lang.Integer> storage,
               VocabCache<VocabWord> vocabCache)

Parameters:

storage - AbstractStorage implementation, key has to be Integer, index of vocabWords

vocabCache - VocabCache implementation, which will be used to lookup word indexes

Method Detail

setCachePerDevice

public StaticWord2Vec.Builder setCachePerDevice(long bytes)

This method lets you to define if decompressed values will be cached, to avoid excessive decompressions. If bytes == 0 - no cache will be used.

Parameters:

bytes -

Returns:

build

public StaticWord2Vec build()

This method returns Static Word2Vec implementation, which is suitable for tasks like neural nets feeding.

Returns: