| Package | Description |
|---|---|
| nuroko.algo | |
| nuroko.core | |
| nuroko.module | |
| nuroko.module.layers | |
| nuroko.module.loss |
| Modifier and Type | Method and Description |
|---|---|
static void |
SimpleBackProp.train(IComponent comp,
mikera.vectorz.AVector input,
mikera.vectorz.AVector target,
double learnRate) |
void |
MomentumBackProp.train(IComponent comp,
mikera.vectorz.AVector input,
mikera.vectorz.AVector target,
double learnRate) |
static void |
SimpleBackProp.train(IComponent comp,
mikera.vectorz.AVector input,
mikera.vectorz.AVector target,
double learnRate,
LossFunction loss) |
void |
MomentumBackProp.train(IComponent comp,
mikera.vectorz.AVector input,
mikera.vectorz.AVector target,
double learnRate,
LossFunction loss) |
void |
MomentumBackProp.trainSynth(IComponent comp,
mikera.vectorz.AVector input,
double learnRate) |
| Constructor and Description |
|---|
MomentumBackProp(IComponent nn,
double momentum) |
| Modifier and Type | Method and Description |
|---|---|
static IComponent |
Components.bias(int length) |
IComponent |
IComponent.clone() |
IComponent |
ISynthesiser.getDownStack() |
IComponent |
ISynthesiser.getUpStack() |
| Modifier and Type | Method and Description |
|---|---|
List<IComponent> |
IComponent.getComponents() |
| Modifier and Type | Method and Description |
|---|---|
void |
IConstraint.applyTo(IComponent c) |
static ALayerStack |
Components.asLayerStack(IComponent comp) |
static Join |
Components.join(IComponent... components) |
static Stack |
Components.stack(IComponent... components) |
| Modifier and Type | Method and Description |
|---|---|
static Join |
Components.join(List<? extends IComponent> components) |
static Stack |
Components.stack(List<? extends IComponent> components) |
| Modifier and Type | Class and Description |
|---|---|
class |
AComponent |
class |
ACompoundComponent
Abstract base class for components that combine other components.
|
class |
ALayerStack
Abstract base class representing a stack of AWeightLayer components
|
class |
AOperationComponent
Base class for components that perform a single operation on each element
from input to output
|
class |
AWeightLayer
Abstract base class representing a weighted linear transformation.
|
class |
Bias |
class |
CompoundLayerStack |
class |
Dropout
Component that performs dropout on activation values.
|
class |
GradientAmplifier |
class |
Identity |
class |
Join
Class representing a side-by-side joining of components
|
class |
NeuralNet |
class |
Normaliser |
class |
Operator |
class |
Sparsifier |
class |
Stack
A stack of connected components, with the output of each feeding into the
input of the next
|
class |
ThinkingOp |
class |
TrainingOp |
| Modifier and Type | Field and Description |
|---|---|
protected ArrayList<IComponent> |
ACompoundComponent.components |
| Modifier and Type | Method and Description |
|---|---|
IComponent |
ACompoundComponent.getComponent(int i) |
IComponent |
AComponent.getComponent(int i) |
IComponent |
Identity.getDownStack() |
IComponent |
AComponent.getDownStack() |
IComponent |
AComponent.getUpStack() |
IComponent |
Stack.topComponent() |
IComponent |
AComponent.topComponent() |
| Modifier and Type | Method and Description |
|---|---|
List<IComponent> |
Normaliser.getComponents() |
List<IComponent> |
NeuralNet.getComponents() |
List<IComponent> |
CompoundLayerStack.getComponents() |
List<IComponent> |
AOperationComponent.getComponents() |
List<IComponent> |
ACompoundComponent.getComponents() |
Iterator<IComponent> |
AComponent.iterator() |
| Constructor and Description |
|---|
ACompoundComponent(List<? extends IComponent> comps) |
Join(List<? extends IComponent> comps) |
Stack(List<? extends IComponent> comps) |
| Modifier and Type | Class and Description |
|---|---|
class |
FullWeightLayer
Fully connected weight layer
|
class |
SparseWeightLayer
Fully connected weight layer
|
| Modifier and Type | Method and Description |
|---|---|
void |
WeightLengthConstraint.applyTo(IComponent c) |
abstract void |
AConstraint.applyTo(IComponent c) |
| Modifier and Type | Method and Description |
|---|---|
void |
LossFunction.calculateErrorDerivative(mikera.vectorz.AVector output,
mikera.vectorz.AVector target,
IComponent comp)
Calculates an error derivative, storing it in the output gradient of the given component
|
Copyright © 2013. All Rights Reserved.