Class MLPExamples

java.lang.Object

org.tribuo.interop.tensorflow.example.MLPExamples

public abstract class MLPExamples
extends Object

Static factory methods which produce Multi-Layer Perceptron architectures.

Method Summary

Modifier and Type	Method	Description
static GraphDefTuple	buildMLPGraph(String inputName, int numFeatures, int[] hiddenSizes, int numOutputs)	Builds an MLP which expects the supplied number of inputs, has hiddenSizes.length hidden layers, before emitting numOutput outputs.

Methods inherited from class java.lang.Object

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

Method Details

buildMLPGraph

public static GraphDefTuple buildMLPGraph(String inputName,
 int numFeatures,
 int[] hiddenSizes,
 int numOutputs)

Builds an MLP which expects the supplied number of inputs, has hiddenSizes.length hidden layers, before emitting numOutput outputs. Uses ReLU as the activation function for the hidden layers.

Parameters:

inputName - The name of the input placeholder.

numFeatures - The number of input features.

hiddenSizes - The hidden layer sizes.

numOutputs - The number of output dimensions.

Returns:

A pair of a graph and the name of the output operation.