org.ea.javacnn

Class JavaCNN

java.lang.Object

org.ea.javacnn.JavaCNN

All Implemented Interfaces:

java.io.Serializable

public class JavaCNN
extends java.lang.Object
implements java.io.Serializable

A network class holding the layers and some helper functions for training and validation.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
JavaCNN(java.util.List<Layer> layers)

Method Summary

Modifier and Type	Method and Description
double	backward(int y) Backprop: compute gradients wrt all parameters
DataBlock	forward(DataBlock db, boolean training)
java.util.List<BackPropResult>	getBackPropagationResult() Accumulate parameters and gradients for the entire network
double	getCostLoss(DataBlock db, int y)
int	getPrediction() This is a convenience function for returning the argmax prediction, assuming the last layer of the net is a softmax
java.lang.String	getPredictions(int[] correctPredictions, int[] numberDistribution, int totalSize, int numOfClasses) Print prediction.
static JavaCNN	loadModel(java.lang.String fileName) Load the model of this neural network from a file.
void	saveModel(java.lang.String fileName) Save convolutional network to a file.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

JavaCNN

public JavaCNN(java.util.List<Layer> layers)

Method Detail

forward

public DataBlock forward(DataBlock db,
                         boolean training)

getCostLoss

public double getCostLoss(DataBlock db,
                          int y)

backward

public double backward(int y)

Backprop: compute gradients wrt all parameters

getBackPropagationResult

public java.util.List<BackPropResult> getBackPropagationResult()

Accumulate parameters and gradients for the entire network

getPrediction

public int getPrediction()

This is a convenience function for returning the argmax prediction, assuming the last layer of the net is a softmax

getPredictions

public java.lang.String getPredictions(int[] correctPredictions,
                                       int[] numberDistribution,
                                       int totalSize,
                                       int numOfClasses)

Print prediction.

saveModel

public void saveModel(java.lang.String fileName)

Save convolutional network to a file.

Parameters:

fileName - output file name.

loadModel

public static JavaCNN loadModel(java.lang.String fileName)

Load the model of this neural network from a file.

Parameters:

fileName - input file name

Returns: