deepnetts.net.layers

Class FlattenLayer

java.lang.Object

deepnetts.net.layers.AbstractLayer

deepnetts.net.layers.FlattenLayer

All Implemented Interfaces:

Layer, Serializable

public class FlattenLayer
extends AbstractLayer

Just transform outputs from prev 3d layer into flat 1d tensor in formward pass backward pass will be same as backward for fc, and for 3d layers backwardFromFc Automaticly add after 2d or 3d layer to transition to fc layers. Simplifiees implementation Now Convolutional and MaxPooling and 3d input layers should recognize flatten layers Treba sam sa ih automatski ubacuje posle conc i maxpooling layera a pre fc layera !!! Kako da se inicijalizuje, poveze i kako radi backward pass - - ovo je glavno sto treba resiti

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
FlattenLayer()

Method Summary

Modifier and Type	Method and Description
void	applyWeightChanges() Applies weight changes to current weights Must be diferent for convolutional does nothing for MaxPooling Same for FullyConnected and OutputLayer
void	backward() This method should implement backward pass in subclasses
void	forward() This method should implement forward pass in subclasses
void	init() This method should implement layer initialization in subclasses, when a layer is added to the network (create weights, outputs, deltas, randomization etc.).
void	setOptimizerType(OptimizerType optType)
String	toString()

Methods inherited from class deepnetts.net.layers.AbstractLayer

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

Methods inherited from class java.lang.Object

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

Constructor Detail

FlattenLayer

public FlattenLayer()

Method Detail

init

public void init()

Description copied from class: AbstractLayer

This method should implement layer initialization in subclasses, when a layer is added to the network (create weights, outputs, deltas, randomization etc.).

Specified by:

init in class AbstractLayer

forward

public void forward()

Description copied from class: AbstractLayer

This method should implement forward pass in subclasses

Specified by:

forward in interface Layer

Specified by:

forward in class AbstractLayer

backward

public void backward()

Description copied from class: AbstractLayer

This method should implement backward pass in subclasses

Specified by:

backward in interface Layer

Specified by:

backward in class AbstractLayer

setOptimizerType

public void setOptimizerType(OptimizerType optType)

Overrides:

setOptimizerType in class AbstractLayer

applyWeightChanges

public void applyWeightChanges()

Description copied from class: AbstractLayer

Applies weight changes to current weights Must be diferent for convolutional does nothing for MaxPooling Same for FullyConnected and OutputLayer

Specified by:

applyWeightChanges in class AbstractLayer

toString

public String toString()

Overrides:

toString in class Object