ConvexOptimizer (deeplearning4j-nn 1.0.0-beta API)

All Superinterfaces:

Serializable

All Known Implementing Classes:

BaseOptimizer, ConjugateGradient, LBFGS, LineGradientDescent, StochasticGradientDescent
```
public interface ConvexOptimizer
extends Serializable
```
Convex optimizer.

Author:

Adam Gibson

Method Summary

All Methods Instance Methods Abstract Methods
Modifier and Type	Method and Description
`int`	`batchSize()` The batch size for the optimizer
`boolean`	`checkTerminalConditions(org.nd4j.linalg.api.ndarray.INDArray gradient, double oldScore, double score, int iteration)` Check termination conditions setup a search state
`ComputationGraphUpdater`	`getComputationGraphUpdater()`
`NeuralNetConfiguration`	`getConf()`
`GradientsAccumulator`	`getGradientsAccumulator()` This method returns GradientsAccumulator instance used in this optimizer.
`StepFunction`	`getStepFunction()` This method returns StepFunction defined within this Optimizer instance
`Updater`	`getUpdater()`
`org.nd4j.linalg.primitives.Pair<Gradient,Double>`	`gradientAndScore(LayerWorkspaceMgr workspaceMgr)` The gradient and score for this optimizer
`boolean`	`optimize(LayerWorkspaceMgr workspaceMgr)` Calls optimize
`void`	`postStep(org.nd4j.linalg.api.ndarray.INDArray line)` After the step has been made, do an action
`void`	`preProcessLine()` Pre preProcess a line before an iteration
`double`	`score()` The score for the optimizer so far
`void`	`setBatchSize(int batchSize)` Set the batch size for the optimizer
`void`	`setGradientsAccumulator(GradientsAccumulator accumulator)` This method specifies GradientsAccumulator instance to be used for updates sharing across multiple models
`void`	`setListeners(Collection<TrainingListener> listeners)`
`void`	`setUpdater(Updater updater)`
`void`	`setUpdaterComputationGraph(ComputationGraphUpdater updater)`
`void`	`setupSearchState(org.nd4j.linalg.primitives.Pair<Gradient,Double> pair)` Based on the gradient and score setup a search state
`void`	`updateGradientAccordingToParams(Gradient gradient, Model model, int batchSize, LayerWorkspaceMgr workspaceMgr)` Update the gradient according to the configuration such as adagrad, momentum, and sparsity

- Method Detail
  - score
```
double score()
```
    The score for the optimizer so far
    
    Returns:
    
    the score for this optimizer so far
  - getUpdater
```
Updater getUpdater()
```
  - getComputationGraphUpdater
```
ComputationGraphUpdater getComputationGraphUpdater()
```
  - setUpdater
```
void setUpdater(Updater updater)
```
  - setUpdaterComputationGraph
```
void setUpdaterComputationGraph(ComputationGraphUpdater updater)
```
  - setListeners
```
void setListeners(Collection<TrainingListener> listeners)
```
  - setGradientsAccumulator
```
void setGradientsAccumulator(GradientsAccumulator accumulator)
```
    This method specifies GradientsAccumulator instance to be used for updates sharing across multiple models
    
    Parameters:
    
    accumulator -
  - getStepFunction
```
StepFunction getStepFunction()
```
    This method returns StepFunction defined within this Optimizer instance
    
    Returns:
  - getGradientsAccumulator
```
GradientsAccumulator getGradientsAccumulator()
```
    This method returns GradientsAccumulator instance used in this optimizer. This method can return null.
    
    Returns:
  - getConf
```
NeuralNetConfiguration getConf()
```
  - gradientAndScore
```
org.nd4j.linalg.primitives.Pair<Gradient,Double> gradientAndScore(LayerWorkspaceMgr workspaceMgr)
```
    The gradient and score for this optimizer
    
    Returns:
    
    the gradient and score for this optimizer
  - optimize
```
boolean optimize(LayerWorkspaceMgr workspaceMgr)
```
    Calls optimize
    
    Returns:
    
    whether the convex optimizer converted or not
  - batchSize
```
int batchSize()
```
    The batch size for the optimizer
    
    Returns:
  - setBatchSize
```
void setBatchSize(int batchSize)
```
    Set the batch size for the optimizer
    
    Parameters:
    
    batchSize -
  - preProcessLine
```
void preProcessLine()
```
    Pre preProcess a line before an iteration
  - postStep
```
void postStep(org.nd4j.linalg.api.ndarray.INDArray line)
```
    After the step has been made, do an action
    
    Parameters:
    
    line -
  - setupSearchState
```
void setupSearchState(org.nd4j.linalg.primitives.Pair<Gradient,Double> pair)
```
    Based on the gradient and score setup a search state
    
    Parameters:
    
    pair - the gradient and score
  - updateGradientAccordingToParams
```
void updateGradientAccordingToParams(Gradient gradient,
                                     Model model,
                                     int batchSize,
                                     LayerWorkspaceMgr workspaceMgr)
```
    Update the gradient according to the configuration such as adagrad, momentum, and sparsity
    
    Parameters:
    
    gradient - the gradient to modify
    
    model - the model with the parameters to update
    
    batchSize - batchSize for update
  - checkTerminalConditions
```
boolean checkTerminalConditions(org.nd4j.linalg.api.ndarray.INDArray gradient,
                                double oldScore,
                                double score,
                                int iteration)
```
    Check termination conditions setup a search state
    
    Parameters:
    
    gradient - layer gradients
    
    iteration - what iteration the optimizer is on

Interface ConvexOptimizer

Method Summary

Method Detail

score

getUpdater

getComputationGraphUpdater

setUpdater

setUpdaterComputationGraph

setListeners

setGradientsAccumulator

getStepFunction

getGradientsAccumulator

getConf

gradientAndScore

optimize

batchSize

setBatchSize

preProcessLine

postStep

setupSearchState

updateGradientAccordingToParams

checkTerminalConditions