net.scala

    var net = FeedForward.create(neuralPrefs)
    val sparkNetwork = new SparkDl4jMultiLayer(sc, net)

    val testIter: DataSetIterator = new FeedForwardIterator(...)
    val trainIter: List[DataSet] = new FeedForwardIterator(...).asScala.toList
    val rddTrain: JavaRDD[DataSet] = sc.parallelize(trainIter)

    for (i <- 0 until neuralPrefs.epochs) {
      net = sparkNetwork.fitDataSet(rddTrain, 200, 2)
      val eval = NeuralEvaluation(net, testIter, i, label)
      eval.log()
      testIter.reset()
    }

    net