Package me.yixqiao.jlearn.models

Class Model.FitBuilder

java.lang.Object

me.yixqiao.jlearn.models.Model.FitBuilder

Enclosing class:

Model

public static class Model.FitBuilder
extends java.lang.Object

Builder class for fit operation.

Field Summary

Fields

Modifier and Type	Field	Description
protected int	batchSize	Batch size.
protected int	epochs	Number of epochs.
protected Matrix	evalX	Evaluation input.
protected Matrix	evalY	Evaluation correct values.
protected boolean	hasEval	Whether to use an evaluation dataset.
protected int	logInterval	Log to console every n epochs.
protected java.util.ArrayList<Metric>	metrics	Metrics to gauge model performance.
protected Optimizer	optimizer	Optimizer.
protected Matrix	trainX	Training input.
protected Matrix	trainY	Training correct values.

Constructor Summary

Constructors

Constructor	Description
FitBuilder(Matrix trainX, Matrix trainY)	Create a new FitBuilder.

Method Summary

Modifier and Type	Method	Description
Model.FitBuilder	batchSize(int batchSize)	Set batch size.
Model.FitBuilder	epochs(int epochs)	Set the epoch number.
Model.FitBuilder	eval(Matrix evalX, Matrix evalY)	Set evaluation dataset.
Model.FitBuilder	logInterval(int logInterval)	Set the log interval.
Model.FitBuilder	metrics(java.util.ArrayList<Metric> metrics)	Set metrics.
Model.FitBuilder	optimizer(Optimizer optimizer)	Set the optimizer.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

trainX

protected Matrix trainX

Training input.

trainY

protected Matrix trainY

Training correct values.

optimizer

protected Optimizer optimizer

Optimizer.

batchSize

protected int batchSize

Batch size.

epochs

protected int epochs

Number of epochs.

metrics

protected java.util.ArrayList<Metric> metrics

Metrics to gauge model performance.

hasEval

protected boolean hasEval

Whether to use an evaluation dataset.

evalX

protected Matrix evalX

Evaluation input.

evalY

protected Matrix evalY

Evaluation correct values.

logInterval

protected int logInterval

Log to console every n epochs.

Constructor Detail

FitBuilder

public FitBuilder(Matrix trainX,
                  Matrix trainY)

Create a new FitBuilder.

Parameters:

trainX - training input

trainY - training correct values

Method Detail

optimizer

public Model.FitBuilder optimizer(Optimizer optimizer)

Set the optimizer.

Parameters:

optimizer - optimizer

Returns:

the instance for daisy chaining

batchSize

public Model.FitBuilder batchSize(int batchSize)

Set batch size.

Parameters:

batchSize - batch size

Returns:

the instance for daisy chaining

epochs

public Model.FitBuilder epochs(int epochs)

Set the epoch number.

Parameters:

epochs - number of epochs

Returns:

the instance for daisy chaining

metrics

public Model.FitBuilder metrics(java.util.ArrayList<Metric> metrics)

Set metrics.

Parameters:

metrics - list of metrics

Returns:

the instance for daisy chaining

eval

public Model.FitBuilder eval(Matrix evalX,
                             Matrix evalY)

Set evaluation dataset.

Parameters:

evalX - evaluation input

evalY - evaluation correct values

Returns:

the instance for daisy chaining

logInterval

public Model.FitBuilder logInterval(int logInterval)

Set the log interval.

Parameters:

logInterval - interval to log

Returns:

the instance for daisy chaining