Java example source code file (Word2VecPerformer.java)

This example Java source code file (Word2VecPerformer.java) is included in the alvinalexander.com "Java Source Code Warehouse" project. The intent of this project is to help you "Learn Java by Example" ^TM.

Learn more about this Java project at its project page.

Java - Java tags/keywords

atomiclong, broadcast, bytearrayinputstream, datainputstream, exception, indarray, inmemorylookuptable, list, logger, max_exp, override, util, vocabword, voidfunction, word2vecperformer

The Word2VecPerformer.java Java example source code

/*
 *
 *  * Copyright 2015 Skymind,Inc.
 *  *
 *  *    Licensed under the Apache License, Version 2.0 (the "License");
 *  *    you may not use this file except in compliance with the License.
 *  *    You may obtain a copy of the License at
 *  *
 *  *        http://www.apache.org/licenses/LICENSE-2.0
 *  *
 *  *    Unless required by applicable law or agreed to in writing, software
 *  *    distributed under the License is distributed on an "AS IS" BASIS,
 *  *    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 *  *    See the License for the specific language governing permissions and
 *  *    limitations under the License.
 *
 */

package org.deeplearning4j.spark.models.embeddings.word2vec;

import org.apache.commons.math3.util.FastMath;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.function.VoidFunction;
import org.apache.spark.broadcast.Broadcast;
import org.deeplearning4j.berkeley.Pair;
import org.deeplearning4j.models.embeddings.inmemory.InMemoryLookupTable;
import org.deeplearning4j.models.word2vec.VocabWord;
import org.nd4j.linalg.api.buffer.DataBuffer;
import org.nd4j.linalg.api.ndarray.INDArray;
import org.nd4j.linalg.factory.Nd4j;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.ByteArrayInputStream;
import java.io.DataInputStream;
import java.io.IOException;
import java.util.List;
import java.util.concurrent.atomic.AtomicLong;

/**
 * Base line word 2 vec performer
 *
 * @author Adam Gibson
 */
@Deprecated
public class Word2VecPerformer implements VoidFunction<Pair> {

    private static double MAX_EXP = 6;
    private boolean useAdaGrad = false;
    private double negative = 5;
    private int numWords = 1;
    private INDArray table;
    private int window = 5;
    private AtomicLong nextRandom = new AtomicLong(5);
    private double alpha = 0.025;
    private double minAlpha = 1e-2;
    private int totalWords = 1;
    private static transient final Logger log = LoggerFactory.getLogger(Word2VecPerformer.class);
    private int lastChecked = 0;
    private  Broadcast<AtomicLong> wordCount;
    private InMemoryLookupTable weights;
    private double[] expTable = new double[1000];
    private int vectorLength;


    public Word2VecPerformer(SparkConf sc, Broadcast<AtomicLong> wordCount, InMemoryLookupTable weights) {
        this.weights = weights;
        this.wordCount = wordCount;
        setup(sc);
    }

    public void setup(SparkConf conf) {
        useAdaGrad = conf.getBoolean(Word2VecVariables.ADAGRAD, false);
        negative = conf.getDouble(Word2VecVariables.NEGATIVE, 5);
        numWords = conf.getInt(Word2VecVariables.NUM_WORDS, 1);
        window = conf.getInt(Word2VecVariables.WINDOW, 5);
        alpha = conf.getDouble(Word2VecVariables.ALPHA, 0.025f);
        minAlpha = conf.getDouble(Word2VecVariables.MIN_ALPHA, 1e-2f);
        totalWords = conf.getInt(Word2VecVariables.NUM_WORDS, 1);
        vectorLength = conf.getInt(Word2VecVariables.VECTOR_LENGTH,100);
        initExpTable();

        if(negative > 0 && conf.contains(Word2VecVariables.TABLE)) {
            try {
                ByteArrayInputStream bis = new ByteArrayInputStream(conf.get(Word2VecVariables.TABLE).getBytes());
                DataInputStream dis = new DataInputStream(bis);
                table = Nd4j.read(dis);
            } catch (IOException e) {
                e.printStackTrace();
            }

        }

    }



    /**
     * Train on a list of vocab words
     * @param sentence the list of vocab words to train on
     */
    public void trainSentence(final List<VocabWord> sentence,double alpha) {
        if (sentence != null && !sentence.isEmpty()) {
            for (int i = 0; i < sentence.size(); i++) {
                if (!sentence.get(i).getWord().endsWith("STOP")) {
                    nextRandom.set(nextRandom.get() * 25214903917L + 11);
                    skipGram(i, sentence, (int) nextRandom.get() % window, alpha);
                }
            }
        }

    }


    /**
     * Train via skip gram
     * @param i
     * @param sentence
     */
    public void skipGram(int i,List<VocabWord> sentence, int b,double alpha) {

        final VocabWord word = sentence.get(i);
        if (word != null && !sentence.isEmpty()) {
            int end = window * 2 + 1 - b;
            for (int a = b; a < end; a++) {
                if (a != window) {
                    int c = i - window + a;
                    if (c >= 0 && c < sentence.size()) {
                        VocabWord lastWord = sentence.get(c);
                        iterateSample(word, lastWord, alpha);
                    }
                }
            }
        }
    }



    /**
     * Iterate on the given 2 vocab words
     *
     * @param w1 the first word to iterate on
     * @param w2 the second word to iterate on
     */
    public  void iterateSample(VocabWord w1, VocabWord w2,double alpha) {
        if(w2 == null || w2.getIndex() < 0)
            return;

        //current word vector
        INDArray l1 = weights.vector(w2.getWord());


        //error for current word and context
        INDArray neu1e = Nd4j.create(vectorLength);

        for(int i = 0; i < w1.getCodeLength(); i++) {
            int code = w1.getCodes().get(i);
            int point = w1.getPoints().get(i);

            INDArray syn1 = weights.getSyn1().slice(point);

            double dot = Nd4j.getBlasWrapper().dot(l1,syn1);

            if (dot >= -MAX_EXP && dot < MAX_EXP) {

                int idx = (int) ((dot + MAX_EXP) * ((double) expTable.length / MAX_EXP / 2.0));
                if (idx >= expTable.length)
                    continue;

                //score
                double f = expTable[idx];
                //gradient
                double g = (1 - code - f) * (useAdaGrad ? w1.getGradient(i, alpha, this.alpha) : alpha);

                Nd4j.getBlasWrapper().level1().axpy(l1.length(), g, syn1, neu1e);
                Nd4j.getBlasWrapper().level1().axpy(l1.length(), g, l1, syn1);
            }


        }


        //negative sampling
        if(negative > 0) {
            int target = w1.getIndex();
            int label;
            INDArray syn1Neg = weights.getSyn1Neg().slice(target);

            for (int d = 0; d < negative + 1; d++) {
                if (d == 0) {

                    label = 1;
                } else {
                    nextRandom.set(nextRandom.get() * 25214903917L + 11);
                    target = table.getInt((int) (nextRandom.get() >> 16) % table.length());
                    if (target == 0)
                        target = (int) nextRandom.get() % (numWords - 1) + 1;
                    if (target == w1.getIndex())
                        continue;
                    label = 0;
                }

                double f = Nd4j.getBlasWrapper().dot(l1, syn1Neg);
                double g;
                if (f > MAX_EXP)
                    g = useAdaGrad ? w1.getGradient(target, (label - 1), this.alpha) : (label - 1) *  alpha;
                else if (f < -MAX_EXP)
                    g = label * (useAdaGrad ?  w1.getGradient(target, alpha, this.alpha) : alpha);
                else
                    g = useAdaGrad ? w1.getGradient(target, label - expTable[(int)((f + MAX_EXP) * (expTable.length / MAX_EXP / 2))], this.alpha) : (label - expTable[(int)((f + MAX_EXP) * (expTable.length / MAX_EXP / 2))]) *   alpha;
                if(syn1Neg.data().dataType() == DataBuffer.Type.DOUBLE)
                    Nd4j.getBlasWrapper().axpy(g,neu1e,l1);
                else
                    Nd4j.getBlasWrapper().axpy((float) g,neu1e,l1);

                if(syn1Neg.data().dataType() == DataBuffer.Type.DOUBLE)
                    Nd4j.getBlasWrapper().axpy(g,syn1Neg,l1);
                else
                    Nd4j.getBlasWrapper().axpy((float) g,syn1Neg,l1);
            }
        }

        if(neu1e.data().dataType() == DataBuffer.Type.DOUBLE)
            Nd4j.getBlasWrapper().axpy(1.0,neu1e,l1);

        else
            Nd4j.getBlasWrapper().axpy(1.0f,neu1e,l1);

    }

    private void initExpTable() {
        for (int i = 0; i < expTable.length; i++) {
            double tmp =   FastMath.exp((i / (double) expTable.length * 2 - 1) * MAX_EXP);
            expTable[i]  = tmp / (tmp + 1.0);
        }
    }


    @Override
    public void call(Pair<List pair) throws Exception {
        double numWordsSoFar = wordCount.getValue().doubleValue();

        List<VocabWord> sentence = pair.getFirst();
        double alpha2 = Math.max(minAlpha, alpha * (1 - (1.0 * numWordsSoFar / (double) totalWords)));
        int totalNewWords = 0;
        trainSentence(sentence, alpha2);
        totalNewWords += sentence.size();



        double newWords = totalNewWords + numWordsSoFar;
        double diff = Math.abs(newWords - lastChecked);
        if(diff >= 10000) {
            lastChecked = (int) newWords;
            log.info("Words so far " + newWords + " out of " + totalWords);
        }

        pair.getSecond().getAndAdd((long) totalNewWords);
    }


}