// some common parameters
			NeuralNetConfiguration.Builder builder = new NeuralNetConfiguration.Builder();
			builder.seed(123);
			builder.biasInit(0);
			builder.miniBatch(false);
			builder.updater(new RmsProp(0.001));
			builder.weightInit(WeightInit.XAVIER);

			ListBuilder listBuilder = builder.list();

			// first difference, for rnns we need to use GravesLSTM.Builder
			for (int i = 0; i < HIDDEN_LAYER_CONT; i++) {
				GravesLSTM.Builder hiddenLayerBuilder = new GravesLSTM.Builder();
				hiddenLayerBuilder.nIn(i == 0 ? LEARNSTRING_CHARS.size() : HIDDEN_LAYER_WIDTH);
				hiddenLayerBuilder.nOut(HIDDEN_LAYER_WIDTH);
				// adopted activation function from GravesLSTMCharModellingExample
				// seems to work well with RNNs
				hiddenLayerBuilder.activation(Activation.TANH);
				listBuilder.layer(i, hiddenLayerBuilder.build());
			}

			// we need to use RnnOutputLayer for our RNN
			RnnOutputLayer.Builder outputLayerBuilder = new RnnOutputLayer.Builder(LossFunction.MCXENT);
			// softmax normalizes the output neurons, the sum of all outputs is 1
			// this is required for our sampleFromDistribution-function
			outputLayerBuilder.activation(Activation.SOFTMAX);
			outputLayerBuilder.nIn(HIDDEN_LAYER_WIDTH);
			outputLayerBuilder.nOut(LEARNSTRING_CHARS.size());
			listBuilder.layer(HIDDEN_LAYER_CONT, outputLayerBuilder.build());

			// finish builder
			listBuilder.pretrain(false);
			listBuilder.backprop(true);

			// create network
			MultiLayerConfiguration conf = listBuilder.build();
			net = new MultiLayerNetwork(conf);

			net.init();
			//net.setListeners(new ScoreIterationListener(1));

///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
      // some epochs
			for (int epoch = 0; epoch < 1000; epoch++) {

				System.out.println("Epoch " + epoch);

				provideUIServer();

				// train the data
				net.fit(trainingData);
				System.out.println("batch " + net.batchSize());

        // clear current stance from the last example
				net.rnnClearPreviousState();

			}