I am going to start training my dataset with RNN. Of course, I need to input data in 3-D dimension.
So, I need to read my data from csv file right for this.
I tried to do it like in some guides I found and did next:
SequenceRecordReader reader = new CSVSequenceRecordReader(); reader.initialize(new FileSplit(new File("test.csv"))); DataSetIterator trainIter = new SequenceRecordReaderDataSetIterator(reader, miniBatchSize, 6, 7, false);
When I check it in outputs the format :
So, as I see it is readed well how I suppose.
Dataset has 7 features and 1 label.
When I starting train my RNN see like:
MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder() .seed(123) //Random number generator seed for improved repeatability. Optional. .weightInit(WeightInit.XAVIER) .updater(new Nadam()) .gradientNormalization(GradientNormalization.ClipElementWiseAbsoluteValue) //Not always required, but helps with this data set .gradientNormalizationThreshold(0.5) .list() .layer(new LSTM.Builder().activation(Activation.TANH).nIn(numInputs).nOut(10).build()) .layer(new RnnOutputLayer.Builder(LossFunctions.LossFunction.MCXENT) .activation(Activation.SOFTMAX).nIn(10).nOut(numOutputs).build()) .build();
Nothing happens after.
What could be the reason of it ?