Package deepnetts.net.layers

Class InputLayer

java.lang.Object

deepnetts.net.layers.AbstractLayer

deepnetts.net.layers.InputLayer

All Implemented Interfaces:

Layer, Serializable

public class InputLayer
extends AbstractLayer

Input layer in a neural network. It is always the first layer in the network. It accepts external input, and sends it for processing to the next layer in a network. Inputs are given as tensors of float values.

Author:

Zoran Sevarac

See Also:

• AbstractLayer
• Serialized Form

Constructor Summary

Constructors

Constructor	Description
InputLayer(int width)	Creates input layer with specified width, and with height and depth equals to one.
InputLayer(int width, int height)	Creates input layer with specified width and height, and depth=1 (single depth/channel).
InputLayer(int depth, int width, int height)	Creates input layer with specified width, height, and depth (number of channels).

Method Summary

Modifier and Type	Method	Description
void	applyWeightChanges()	This method does nothing in input layer.
void	backward()	This method does nothing in the input layer, and should never be called.
void	forward()	This method does nothing in the input layer, and should never be called.
int	getTensorDim()
final void	init()	Initialize this layer in network.
void	setInput(TensorBase in)	Sets network input
String	toString()

Methods inherited from class deepnetts.net.layers.AbstractLayer

getActivation, getActivationType, getBatchSize, getBiases, getDeltaBiases, getDeltas, getDeltaWeights, getDepth, getForwardAccelerator, getGradients, getHeight, getL1Regularization, getL1WeightSum, getL2Regularization, getL2WeightSum, getLearningRate, getMode, getMomentum, getNextLayer, getOptimizer, getOptimizerType, getOutputs, getPrevDeltaBiases, getPrevDeltaWeights, getPrevlayer, getWeights, getWidth, initTransientFields, isBatchMode, isTrainable, setBatchMode, setBatchSize, setBiases, setCudaHandles, setDeltas, setL1Regularization, setL2Regularization, setLearningRate, setMode, setMomentum, setNextlayer, setOptimizerType, setOutputs, setPrevDeltaWeights, setPrevLayer, setThreadPool, setTrainable, setWeights, setWeights

Methods inherited from class java.lang.Object

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

Constructor Details

InputLayer

public InputLayer(int depth,
 int width,
 int height)

Creates input layer with specified width, height, and depth (number of channels).

Parameters:

depth - layer depth (number of input channels)

width - layer width

height - layer height

InputLayer

public InputLayer(int width,
 int height)

Creates input layer with specified width and height, and depth=1 (single depth/channel).

Parameters:

width - layer width

height - layer height

InputLayer

public InputLayer(int width)

Creates input layer with specified width, and with height and depth equals to one.

Parameters:

width - layer width

Method Details

init

public final void init()

Initialize this layer in network.

Specified by:

init in class AbstractLayer

setInput

public void setInput(TensorBase in)

Sets network input

Parameters:

in - input matrix/array

forward

public void forward()

This method does nothing in the input layer, and should never be called.

Specified by:

forward in interface Layer

Specified by:

forward in class AbstractLayer

backward

public void backward()

This method does nothing in the input layer, and should never be called.

Specified by:

backward in interface Layer

Specified by:

backward in class AbstractLayer

applyWeightChanges

public void applyWeightChanges()

This method does nothing in input layer.

Specified by:

applyWeightChanges in class AbstractLayer

toString

public String toString()

Overrides:

toString in class Object

getTensorDim

public int getTensorDim()