edu.cmu.sphinx.trainer

Interface TrainManager

All Superinterfaces:

Configurable

All Known Implementing Classes:

SimpleTrainManager

public interface TrainManager
extends Configurable

Manages inputs and outputs to the other trainer classes.

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	PROP_MAXIMUM_ITERATION The maximum number of iterations.
static java.lang.String	PROP_MINIMUM_IMPROVEMENT The minimum relative improvement of the log likelihood associated with the training data.

Method Summary

Modifier and Type	Method and Description
void	copyModels(java.lang.String context) Copy the model.
void	initializeModels(java.lang.String context) Initializes the acoustic models.
void	saveModels(java.lang.String context) Saves the acoustic models.
void	train() Do the train.
void	trainContextIndependentModels(java.lang.String context) Trains context independent models.

Methods inherited from interface edu.cmu.sphinx.util.props.Configurable

newProperties

Field Detail

PROP_MINIMUM_IMPROVEMENT

@S4Double(defaultValue=0.20000000298023224)
static final java.lang.String PROP_MINIMUM_IMPROVEMENT

The minimum relative improvement of the log likelihood associated with the training data.

See Also:

Constant Field Values

PROP_MAXIMUM_ITERATION

@S4Integer(defaultValue=15)
static final java.lang.String PROP_MAXIMUM_ITERATION

The maximum number of iterations.

See Also:

Constant Field Values

Method Detail

train

void train()

Do the train.

saveModels

void saveModels(java.lang.String context)
         throws java.io.IOException

Saves the acoustic models.

Parameters:

context - the context of this TrainManager

Throws:

java.io.IOException - if an error occurs while loading the data

copyModels

void copyModels(java.lang.String context)
         throws java.io.IOException

Copy the model.

This method copies to model set, possibly to a new location and new format. This is useful if one wants to convert from binary to ascii and vice versa, or from a directory structure to a JAR file. If only one model is used, then name can be null.

Parameters:

context - this TrainManager's context

Throws:

java.io.IOException - if an error occurs while loading the data

initializeModels

void initializeModels(java.lang.String context)
               throws java.io.IOException

Initializes the acoustic models.

Parameters:

context - the context of this TrainManager

Throws:

java.io.IOException - if something went wrong

trainContextIndependentModels

void trainContextIndependentModels(java.lang.String context)
                            throws java.io.IOException

Trains context independent models. If the initialization stage was skipped, it loads models from files, automatically.

Parameters:

context - the context of this train manager.

Throws:

java.io.IOException - if something went wrong