This is a Spark <-> H2O / Sparkling water deep learning prototype.

h2o-sparkling-water-deep-learning.scala

import org.apache.spark.{SparkConf, SparkContext}
import org.apache.spark.h2o.{H2OContext, H2OFrame}
import org.apache.spark.sql.DataFrame

import hex.deeplearning.DeepLearning
import water.app.SparkContextSupport

import hex.deeplearning.DeepLearningParameters
import hex.deeplearning.DeepLearningParameters.Activation
import org.apache.spark.h2o.{DoubleHolder, H2OContext, H2OFrame}
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.SQLContext
import org.apache.spark.{SparkConf, SparkContext, SparkFiles}
import water.app.SparkContextSupport
import water.fvec.Frame

object DeepRM {
  def predict(sparkContext: SparkContext, trainingDF: DataFrame, testDF: DataFrame, labelColumn: String): DataFrame = {
    val sqlContext = new SQLContext(sparkContext);
    val hc = H2OContext.getOrCreate(sparkContext)

    val testFrame = hc.asH2OFrame(testDF)
    val trainingFrame = hc.asH2OFrame(trainingDF)

    val dlParams = new DeepLearningParameters()

    dlParams._train = trainingFrame.key
    dlParams._response_column = labelColumn

    dlParams._epochs = 20
    dlParams._activation = Activation.RectifierWithDropout
    dlParams._hidden = Array[Int](100, 100)

    val dl = new DeepLearning(dlParams)
    val dlModel = dl.trainModel.get

    // Predict and add prediction column to the test dataframe
    val prediction: Frame = dlModel.score(testFrame)
    val ret = testFrame.add(prediction)

    val predictedDF = hc.asDataFrame(hc.asH2OFrame(ret))(sqlContext)
    hc.stop()

    predictedDF
  }

  def main(args: Array[String]) {
    val conf = new SparkConf().setMaster("local").setAppName("H2O Test")
    val sc = new SparkContext(conf)
    //sc.addJar("/Users/zoltanctoth/src/sparkling-water/assembly/build/libs/sparkling-water-assembly-1.5.99999-SNAPSHOT-all.jar")
    //sc.addJar("/opt/spark/lib/spark-assembly-1.5.2-hadoop2.6.0.jar")
    val sqlContext = new org.apache.spark.sql.SQLContext(sc)
    val training = sqlContext.read.parquet("data/training2.parquet")
    val test = sqlContext.read.parquet("data/test2.parquet")
    val labelCol = "is_late"

    val prDF = predict(sc, training, test, labelCol)
    println(prDF.show())

    sc.stop()
  }
}