Java example source code file (SparkDl4jLayer.java)

This example Java source code file (SparkDl4jLayer.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

add, feedforwardlayer, illegalstateexception, indarray, javardd, layer, logger, neuralnetconfiguration, number, ran, running, serializable, sparkdl4jlayer, vector

The SparkDl4jLayer.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.impl.layer;

import org.apache.spark.SparkContext;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.broadcast.Broadcast;
import org.apache.spark.mllib.linalg.Matrix;
import org.apache.spark.mllib.linalg.Vector;
import org.apache.spark.mllib.regression.LabeledPoint;
import org.canova.api.records.reader.RecordReader;
import org.deeplearning4j.nn.api.Layer;
import org.deeplearning4j.nn.conf.NeuralNetConfiguration;
import org.deeplearning4j.nn.conf.layers.FeedForwardLayer;
import org.deeplearning4j.nn.layers.factory.LayerFactories;
import org.deeplearning4j.spark.canova.RecordReaderFunction;
import org.deeplearning4j.spark.impl.common.Add;
import org.deeplearning4j.spark.util.MLLibUtil;
import org.nd4j.linalg.api.ndarray.INDArray;
import org.nd4j.linalg.dataset.DataSet;
import org.nd4j.linalg.factory.Nd4j;
import parquet.org.slf4j.Logger;
import parquet.org.slf4j.LoggerFactory;

import java.io.Serializable;

/**
 * Master class for org.deeplearning4j.spark
 * layers
 * @author Adam Gibson
 */
public class SparkDl4jLayer implements Serializable {

    private transient SparkContext sparkContext;
    private transient JavaSparkContext sc;
    private NeuralNetConfiguration conf;
    private Layer layer;
    private Broadcast<INDArray> params;
    private boolean averageEachIteration = false;
    private static Logger log = LoggerFactory.getLogger(SparkDl4jLayer.class);


    public SparkDl4jLayer(SparkContext sparkContext, NeuralNetConfiguration conf) {
        this.sparkContext = sparkContext;
        this.conf = conf.clone();
        sc = new JavaSparkContext(this.sparkContext);
    }

    public SparkDl4jLayer(JavaSparkContext sc, NeuralNetConfiguration conf) {
        this.sc = sc;
        this.conf = conf.clone();
    }

    /**
     * Fit the layer based on the specified org.deeplearning4j.spark context text file
     * @param path the path to the text file
     * @param labelIndex the index of the label
     * @param recordReader the record reader
     * @return the fit layer
     */
    public Layer fit(String path,int labelIndex,RecordReader recordReader) {
        FeedForwardLayer ffLayer = (FeedForwardLayer) conf.getLayer();

        JavaRDD<String> lines = sc.textFile(path);
        // gotta map this to a Matrix/INDArray
        JavaRDD<DataSet> points = lines.map(new RecordReaderFunction(recordReader
                , labelIndex, ffLayer.getNOut()));
        return fitDataSet(points);

    }

    /**
     * Fit the given rdd given the context.
     * This will convert the labeled points
     * to the internal dl4j format and train the model on that
     * @param sc the org.deeplearning4j.spark context
     * @param rdd the rdd to fitDataSet
     * @return the multi layer network that was fitDataSet
     */
    public Layer fit(JavaSparkContext sc,JavaRDD<LabeledPoint> rdd) {
        FeedForwardLayer ffLayer = (FeedForwardLayer) conf.getLayer();
        return fitDataSet(MLLibUtil.fromLabeledPoint(sc, rdd, ffLayer.getNOut()));
    }

    /**
     * Fit a java rdd of dataset
     * @param rdd the rdd to fit
     * @return the fit layer
     */
    public Layer fitDataSet(JavaRDD<DataSet> rdd) {
        int iterations = conf.getNumIterations();
        long count = rdd.count();


        log.info("Running distributed training averaging each iteration " + averageEachIteration + " and " + rdd.partitions().size() + " partitions");
        if(!averageEachIteration) {
            int numParams = LayerFactories.getFactory(conf).initializer().numParams(conf,true);
            final INDArray params = Nd4j.create(1, numParams);
            Layer layer = LayerFactories.getFactory(conf.getLayer()).create(conf,null,0,params, true);
            layer.setBackpropGradientsViewArray(Nd4j.create(1,numParams));
//            final INDArray params = layer.params();
            this.params = sc.broadcast(params);
            log.info("Broadcasting initial parameters of length " + params.length());
            int paramsLength = layer.numParams();
            if(params.length() != paramsLength)
                throw new IllegalStateException("Number of params " + paramsLength + " was not equal to " + params.length());
            JavaRDD<INDArray> results = rdd.sample(true,0.4).mapPartitions(new IterativeReduceFlatMap(conf.toJson(), this.params));
            log.debug("Ran iterative reduce...averaging results now.");
            INDArray newParams = results.fold(Nd4j.zeros(results.first().shape()),new Add());
            newParams.divi(rdd.partitions().size());
            layer.setParams(newParams);
            this.layer = layer;
        }
        else {
            conf.setNumIterations(1);
            int numParams = LayerFactories.getFactory(conf).initializer().numParams(conf,true);
            final INDArray params = Nd4j.create(1, numParams);
            Layer layer = LayerFactories.getFactory(conf.getLayer()).create(conf, null, 0, params, true);
            layer.setBackpropGradientsViewArray(Nd4j.create(1,numParams));
//            final INDArray params = layer.params();
            this.params = sc.broadcast(params);

            for(int i = 0; i < iterations; i++) {
                JavaRDD<INDArray> results = rdd.sample(true,0.3).mapPartitions(new IterativeReduceFlatMap(conf.toJson(), this.params));

                int paramsLength = layer.numParams();
                if(params.length() != paramsLength)
                    throw new IllegalStateException("Number of params " + paramsLength + " was not equal to " + params.length());

                INDArray newParams = results.fold(Nd4j.zeros(results.first().shape()), new Add());
                newParams.divi(rdd.partitions().size());
            }

            layer.setParams(this.params.value().dup());
            this.layer = layer;


        }


        return layer;
    }


    /**
     * Predict the given feature matrix
     * @param features the given feature matrix
     * @return the predictions
     */
    public Matrix predict(Matrix features) {
        return MLLibUtil.toMatrix(layer.activate(MLLibUtil.toMatrix(features)));
    }


    /**
     * Predict the given vector
     * @param point the vector to predict
     * @return the predicted vector
     */
    public Vector predict(Vector point) {
        return MLLibUtil.toVector(layer.activate(MLLibUtil.toVector(point)));
    }


    /**
     * Train a multi layer network
     * @param data the data to train on
     * @param conf the configuration of the network
     * @return the fit multi layer network
     */
    public static Layer train(JavaRDD<LabeledPoint> data,NeuralNetConfiguration conf) {
        SparkDl4jLayer multiLayer = new SparkDl4jLayer(data.context(),conf);
        return multiLayer.fit(new JavaSparkContext(data.context()),data);

    }



}