SgdSpace (deeplearning4j 1.0.0-beta6 API)

Skip navigation links

Prev Class
Next Class

All Classes

Summary:
Nested |
Field |
Constr |
Method

Detail:
Field |
Constr |
Method

java.lang.Object
- org.deeplearning4j.arbiter.optimize.api.AbstractParameterSpace<IUpdater>
- - org.deeplearning4j.arbiter.conf.updater.BaseUpdaterSpace
  - - org.deeplearning4j.arbiter.conf.updater.SgdSpace

All Implemented Interfaces:

ParameterSpace<IUpdater>
```
public class SgdSpace
extends BaseUpdaterSpace
```

- Field Summary
  
  Fields
  Modifier and Type Field and Description
  
  protected ParameterSpace<Double> learningRate
  
  protected ParameterSpace<ISchedule> learningRateSchedule
- Constructor Summary
  
  Constructors
  Constructor and Description
  
  SgdSpace(ParameterSpace<Double> learningRate)
  
  SgdSpace(ParameterSpace<Double> learningRate, ParameterSpace<ISchedule> learningRateSchedule)
- Method Summary
  
  All Methods Instance Methods Concrete Methods
  Modifier and Type Method and Description
  
  IUpdater getValue(double[] parameterValues)
  Generate a candidate given a set of values.
  - Methods inherited from class org.deeplearning4j.arbiter.conf.updater.BaseUpdaterSpace
    collectLeaves, isLeaf, numParameters, setIndices
  - Methods inherited from class org.deeplearning4j.arbiter.optimize.api.AbstractParameterSpace
    getNestedSpaces
  - Methods inherited from class java.lang.Object
    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - learningRate
```
protected ParameterSpace<Double> learningRate
```
  - learningRateSchedule
```
protected ParameterSpace<ISchedule> learningRateSchedule
```
- Constructor Detail
  - SgdSpace
```
public SgdSpace(ParameterSpace<Double> learningRate)
```
  - SgdSpace
```
public SgdSpace(ParameterSpace<Double> learningRate,
                ParameterSpace<ISchedule> learningRateSchedule)
```
- Method Detail
  - getValue
```
public IUpdater getValue(double[] parameterValues)
```
    Description copied from interface: ParameterSpace
    
    Generate a candidate given a set of values. These values are then mapped to a specific candidate, using some mapping function (such as the prior probability distribution)
    
    Parameters:
    
    parameterValues - A set of values, each in the range [0,1], of length ParameterSpace.numParameters()

Skip navigation links

Prev Class
Next Class

All Classes

Summary:
Nested |
Field |
Constr |
Method

Detail:
Field |
Constr |
Method

Copyright © 2020. All rights reserved.