org.deeplearning4j.spark.models.embeddings.glove

Class Glove

java.lang.Object

org.deeplearning4j.spark.models.embeddings.glove.Glove

All Implemented Interfaces:

Serializable

public class Glove
extends Object
implements Serializable

Spark glove

Author:

Adam Gibson

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Glove(boolean symmetric, int windowSize, int iterations)
Glove(String tokenizerFactoryClazz, boolean symmetric, int windowSize, int iterations)

Method Summary

Modifier and Type	Method and Description
Pair<VocabCache<VocabWord>,GloveWeightLookupTable>	train(org.apache.spark.api.java.JavaRDD<String> rdd) Train on the corpus

Methods inherited from class java.lang.Object

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

Constructor Detail

Glove

public Glove(String tokenizerFactoryClazz,
             boolean symmetric,
             int windowSize,
             int iterations)

Parameters:

tokenizerFactoryClazz - the fully qualified class name of the tokenizer

symmetric - whether the co occurrence counts should be symmetric

windowSize - the window size for co occurrence

iterations - the number of iterations

Glove

public Glove(boolean symmetric,
             int windowSize,
             int iterations)

Parameters:

symmetric - whether the co occurrence counts should be symmetric

windowSize - the window size for co occurrence

iterations - the number of iterations

Method Detail

train

public Pair<VocabCache<VocabWord>,GloveWeightLookupTable> train(org.apache.spark.api.java.JavaRDD<String> rdd)
                                                         throws Exception

Train on the corpus

Parameters:

rdd - the rdd to train

Returns:

the vocab and weights

Throws:

Exception