Layer.Builder (deeplearning4j-nn 0.8.0 API)

java.lang.Object
- org.deeplearning4j.nn.conf.layers.Layer.Builder<T>

Direct Known Subclasses:

FeedForwardLayer.Builder, GlobalPoolingLayer.Builder, LocalResponseNormalization.Builder, SubsamplingLayer.BaseSubsamplingBuilder, ZeroPaddingLayer.Builder

Enclosing class:

Layer
```
public abstract static class Layer.Builder<T extends Layer.Builder<T>>
extends Object
```

Field Summary

Fields
Modifier and Type	Field and Description
`protected org.nd4j.linalg.activations.IActivation`	`activationFn`
`protected double`	`adamMeanDecay`
`protected double`	`adamVarDecay`
`protected double`	`biasInit`
`protected double`	`biasLearningRate`
`protected Distribution`	`dist`
`protected double`	`dropOut`
`protected double`	`epsilon`
`protected GradientNormalization`	`gradientNormalization`
`protected double`	`gradientNormalizationThreshold`
`protected double`	`l1`
`protected double`	`l1Bias`
`protected double`	`l2`
`protected double`	`l2Bias`
`protected String`	`layerName`
`protected double`	`learningRate`
`protected LearningRatePolicy`	`learningRatePolicy`
`protected Map<Integer,Double>`	`learningRateSchedule`
`protected double`	`momentum`
`protected Map<Integer,Double>`	`momentumAfter`
`protected double`	`rho`
`protected double`	`rmsDecay`
`protected Updater`	`updater`
`protected WeightInit`	`weightInit`

Constructor Summary

Constructors
Constructor and Description

Builder()

Constructors
Constructor and Description
`Builder()`

Method Summary

All Methods Instance Methods Abstract Methods Concrete Methods Deprecated Methods
Modifier and Type	Method and Description
`T`	`activation(org.nd4j.linalg.activations.Activation activation)`
`T`	`activation(org.nd4j.linalg.activations.IActivation activationFunction)`
`T`	`activation(String activationFunction)` Deprecated. Use `activation(Activation)` or `@activation(IActivation)`
`T`	`adamMeanDecay(double adamMeanDecay)` Mean decay rate for Adam updater.
`T`	`adamVarDecay(double adamVarDecay)` Variance decay rate for Adam updater.
`T`	`biasInit(double biasInit)`
`T`	`biasLearningRate(double biasLearningRate)` Bias learning rate.
`abstract <E extends Layer> E`	`build()`
`T`	`dist(Distribution dist)` Distribution to sample initial weights from.
`T`	`dropOut(double dropOut)` Dropout.
`T`	`epsilon(double epsilon)` Epsilon value for updaters: Adagrad and Adadelta.
`T`	`gradientNormalization(GradientNormalization gradientNormalization)` Gradient normalization strategy.
`T`	`gradientNormalizationThreshold(double threshold)` Threshold for gradient normalization, only used for GradientNormalization.ClipL2PerLayer, GradientNormalization.ClipL2PerParamType, and GradientNormalization.ClipElementWiseAbsoluteValue Not used otherwise. L2 threshold for first two types of clipping, or absolute value threshold for last type of clipping.
`T`	`l1(double l1)` L1 regularization coefficient (weights only).
`T`	`l1Bias(double l1Bias)` L1 regularization coefficient for the bias.
`T`	`l2(double l2)` L2 regularization coefficient (weights only).
`T`	`l2Bias(double l2Bias)` L2 regularization coefficient for the bias.
`T`	`learningRate(double learningRate)` Learning rate.
`T`	`learningRateDecayPolicy(LearningRatePolicy policy)` Learning rate decay policy.
`T`	`learningRateSchedule(Map<Integer,Double> learningRateSchedule)` Learning rate schedule.
`T`	`momentum(double momentum)` Momentum rate.
`T`	`momentumAfter(Map<Integer,Double> momentumAfter)` Momentum schedule.
`T`	`name(String layerName)` Layer name assigns layer string name.
`T`	`rho(double rho)` Ada delta coefficient, rho.
`T`	`rmsDecay(double rmsDecay)` Decay rate for RMSProp.
`T`	`updater(Updater updater)` Gradient updater.
`T`	`weightInit(WeightInit weightInit)` Weight initialization scheme.

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - layerName
```
protected String layerName
```
  - activationFn
```
protected org.nd4j.linalg.activations.IActivation activationFn
```
  - weightInit
```
protected WeightInit weightInit
```
  - biasInit
```
protected double biasInit
```
  - dist
```
protected Distribution dist
```
  - learningRate
```
protected double learningRate
```
  - biasLearningRate
```
protected double biasLearningRate
```
  - learningRateSchedule
```
protected Map<Integer,Double> learningRateSchedule
```
  - momentum
```
protected double momentum
```
  - momentumAfter
```
protected Map<Integer,Double> momentumAfter
```
  - l1
```
protected double l1
```
  - l2
```
protected double l2
```
  - l1Bias
```
protected double l1Bias
```
  - l2Bias
```
protected double l2Bias
```
  - dropOut
```
protected double dropOut
```
  - updater
```
protected Updater updater
```
  - rho
```
protected double rho
```
  - epsilon
```
protected double epsilon
```
  - rmsDecay
```
protected double rmsDecay
```
  - adamMeanDecay
```
protected double adamMeanDecay
```
  - adamVarDecay
```
protected double adamVarDecay
```
  - gradientNormalization
```
protected GradientNormalization gradientNormalization
```
  - gradientNormalizationThreshold
```
protected double gradientNormalizationThreshold
```
  - learningRatePolicy
```
protected LearningRatePolicy learningRatePolicy
```
- Constructor Detail
  - Builder
```
public Builder()
```
- Method Detail
  - name
```
public T name(String layerName)
```
    Layer name assigns layer string name. Allows easier differentiation between layers.
  - activation
```
@Deprecated
public T activation(String activationFunction)
```
    Deprecated. Use activation(Activation) or @activation(IActivation)
    
    Layer activation function. Typical values include:
    "relu" (rectified linear), "tanh", "sigmoid", "softmax", "hardtanh", "leakyrelu", "maxout", "softsign", "softplus"
  - activation
```
public T activation(org.nd4j.linalg.activations.IActivation activationFunction)
```
  - activation
```
public T activation(org.nd4j.linalg.activations.Activation activation)
```
  - weightInit
```
public T weightInit(WeightInit weightInit)
```
    Weight initialization scheme.
    
    See Also:
    
    WeightInit
  - biasInit
```
public T biasInit(double biasInit)
```
  - dist
```
public T dist(Distribution dist)
```
    Distribution to sample initial weights from. Used in conjunction with .weightInit(WeightInit.DISTRIBUTION).
  - learningRate
```
public T learningRate(double learningRate)
```
    Learning rate. Defaults to 1e-1
  - biasLearningRate
```
public T biasLearningRate(double biasLearningRate)
```
    Bias learning rate. Set this to apply a different learning rate to the bias
  - learningRateSchedule
```
public T learningRateSchedule(Map<Integer,Double> learningRateSchedule)
```
    Learning rate schedule. Map of the iteration to the learning rate to apply at that iteration.
  - l1
```
public T l1(double l1)
```
    L1 regularization coefficient (weights only). Use l1Bias(double) to configure the l1 regularization coefficient for the bias.
  - l2
```
public T l2(double l2)
```
    L2 regularization coefficient (weights only). Use l2Bias(double) to configure the l2 regularization coefficient for the bias.
  - l1Bias
```
public T l1Bias(double l1Bias)
```
    L1 regularization coefficient for the bias. Default: 0. See also l1(double)
  - l2Bias
```
public T l2Bias(double l2Bias)
```
    L2 regularization coefficient for the bias. Default: 0. See also l2(double)
  - dropOut
```
public T dropOut(double dropOut)
```
    Dropout. Value is probability of retaining an activation - thus 1.0 is equivalent to no dropout. Note that 0.0 (the default) disables dropout.
  - momentum
```
public T momentum(double momentum)
```
    Momentum rate.
  - momentumAfter
```
public T momentumAfter(Map<Integer,Double> momentumAfter)
```
    Momentum schedule. Map of the iteration to the momentum rate to apply at that iteration.
  - updater
```
public T updater(Updater updater)
```
    Gradient updater. For example, SGD for standard stochastic gradient descent, NESTEROV for Nesterov momentum, RSMPROP for RMSProp, etc.
    
    See Also:
    
    Updater
  - rho
```
public T rho(double rho)
```
    Ada delta coefficient, rho. Only applies if using .updater(Updater.ADADELTA)
    
    Parameters:
    
    rho -
  - rmsDecay
```
public T rmsDecay(double rmsDecay)
```
    Decay rate for RMSProp. Only applies if using .updater(Updater.RMSPROP)
  - epsilon
```
public T epsilon(double epsilon)
```
    Epsilon value for updaters: Adagrad and Adadelta. Only used if using Updater.ADAGRAD or Updater.ADADELTA
    
    Parameters:
    
    epsilon - Epsilon value to use for adagrad and adadelta
  - adamMeanDecay
```
public T adamMeanDecay(double adamMeanDecay)
```
    Mean decay rate for Adam updater. Only applies if using .updater(Updater.ADAM)
  - adamVarDecay
```
public T adamVarDecay(double adamVarDecay)
```
    Variance decay rate for Adam updater. Only applies if using .updater(Updater.ADAM)
  - gradientNormalization
```
public T gradientNormalization(GradientNormalization gradientNormalization)
```
    Gradient normalization strategy. Used to specify gradient renormalization, gradient clipping etc.
    
    Parameters:
    
    gradientNormalization - Type of normalization to use. Defaults to None.
    
    See Also:
    
    GradientNormalization
  - gradientNormalizationThreshold
```
public T gradientNormalizationThreshold(double threshold)
```
    Threshold for gradient normalization, only used for GradientNormalization.ClipL2PerLayer, GradientNormalization.ClipL2PerParamType, and GradientNormalization.ClipElementWiseAbsoluteValue
    Not used otherwise.
    L2 threshold for first two types of clipping, or absolute value threshold for last type of clipping.
  - learningRateDecayPolicy
```
public T learningRateDecayPolicy(LearningRatePolicy policy)
```
    Learning rate decay policy. Used to adapt learning rate based on policy.
    
    Parameters:
    
    policy - Type of policy to use. Defaults to None.
    
    See Also:
    
    GradientNormalization
  - build
```
public abstract <E extends Layer> E build()
```

Class Layer.Builder<T extends Layer.Builder<T>>

Field Summary

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Field Detail

layerName

activationFn

weightInit

biasInit

dist

learningRate

biasLearningRate

learningRateSchedule

momentum

momentumAfter

l1

l2

l1Bias

l2Bias

dropOut

updater

rho

epsilon

rmsDecay

adamMeanDecay

adamVarDecay

gradientNormalization

gradientNormalizationThreshold

learningRatePolicy

Constructor Detail

Builder

Method Detail

name

activation

activation

activation

weightInit

biasInit

dist

learningRate

biasLearningRate

learningRateSchedule

l1

l2

l1Bias

l2Bias

dropOut

momentum

momentumAfter

updater

rho

rmsDecay

epsilon

adamMeanDecay

adamVarDecay

gradientNormalization

gradientNormalizationThreshold

learningRateDecayPolicy

build