Hands-On Deep Learning with Apache Spark

CNN distributed training with Spark and DL4J

Let's get back to the example that has been presented in Chapter 5, Convolutional Neural Networks, Hands-on CNN with Spark, about handwritten digits image classification on the MNIST dataset. For convenience, here's a reminder of the network configuration used there:

val channels = 1 val outputNum = 10 val conf = new NeuralNetConfiguration.Builder()     .seed(seed)     .iterations(iterations)     .regularization(true)     .l2(0.0005)     .learningRate(.01)     .weightInit(WeightInit.XAVIER)     .optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT)     .updater(Updater.NESTEROVS)     .momentum(0.9)     .list     .layer(0, new ConvolutionLayer.Builder(5, 5)         .nIn(channels)         .stride(1, 1)         .nOut(20) .activation(Activation.IDENTITY) ...