gistfile1.txt

        SameDiff sd = SameDiff.create();

        //Properties for MNIST dataset:
        int nIn = 28*28;
        int nOut = 10;

        //Create input and label variables
        SDVariable in = sd.var("input", -1, nIn);                 //Shape: [?, 784] - i.e., minibatch x 784 for MNIST
        SDVariable label = sd.var("label", -1, nOut);             //Shape: [?, 10] - i.e., minibatch x 10 for MNIST
        sd.addAsPlaceHolder("input");
        sd.addAsPlaceHolder("label");

        //Define hidden layer - MLP (fully connected)
        int layerSize0 = 128;
        SDVariable w0 = sd.var("w0", new XavierInitScheme('c', nIn, layerSize0), nIn, layerSize0);
        SDVariable b0 = sd.zero("b0", 1, layerSize0);
        SDVariable activations0 = sd.tanh(in.mmul(w0).add(b0));

        //Define output layer - MLP (fully connected) + softmax
        SDVariable w1 = sd.var("w1", new XavierInitScheme('c', layerSize0, nOut), layerSize0, nOut);
        SDVariable b1 = sd.zero("b1", 1, nOut);

        SDVariable z1 = activations0.mmul(w1).add("prediction", b1);
        SDVariable softmax = sd.softmax("softmax", z1);

        //Define loss function:
        SDVariable diff = sd.f().squaredDifference(softmax, label);
        SDVariable lossMse = diff.mean();

        //Create and set the training configuration
        double learningRate = 1e-3;
        TrainingConfig config = new TrainingConfig.Builder()
            .l2(1e-4)                               //L2 regularization
            .updater(new Adam(learningRate))        //Adam optimizer with specified learning rate
            .dataSetFeatureMapping("input")         //DataSet features array should be associated with variable "input"
            .dataSetLabelMapping("label")           //DataSet label array should be associated with variable "label"
            .build();

        sd.setTrainingConfig(config);

        int batchSize = 32;
        DataSetIterator trainData = new MnistDataSetIterator(batchSize, true, 12345);
        DataSetIterator testData = new MnistDataSetIterator(batchSize, false, 12345);

        //Perform training for 2 epochs
        int numEpochs = 2;
        sd.fit(trainData, numEpochs);

        //Evaluate on test set:
        String outputVariable = "softmax";
        Evaluation evaluation = new Evaluation();
        sd.evaluate(testData, outputVariable, evaluation);

        //Print evaluation statistics:
        System.out.println(evaluation.stats());

        //Save the trained network. 2 options
        //Saving Option 1: Save for inference only - no updater state. FlatBuffers format
        File saveFileForInference = new File("sameDiffExampleInference.fb");
        sd.asFlatFile(saveFileForInference);

        SameDiff loadedForInference = SameDiff.fromFlatFile(saveFileForInference);

        //Saving Option 2: Save for further training - with updater state (larger file size, updater not required for inference). Zip format
        File saveFileForFurtherTraining = new File("sameDiffExampleTraining.zip");
        sd.saveWithTrainingConfig(saveFileForFurtherTraining);

        SameDiff loadedForFurtherTraining = SameDiff.restoreFromTrainingConfigZip(saveFileForFurtherTraining);
        loadedForFurtherTraining.fit(trainData, 1);         //************ EXCEPTION HERE **************
        evaluation = new Evaluation();
        sd.evaluate(testData, outputVariable, evaluation);
        System.out.println(evaluation.stats());