mckunkel

public static ComputationGraph KunkelPetersYolo(int height, int width, int numChannels) {

		int numClasses = 14;
		double[][] priorBoxes = FaultUtils.allPriors;
		int nBoxes = priorBoxes.length;

		INDArray priors = Nd4j.create(priorBoxes);
		double lambdaNoObj = 0.5;
		double lambdaCoord = 1.0;
		GraphBuilder graphBuilder = new NeuralNetConfiguration.Builder().weightInit(WeightInit.XAVIER)

mckunkel

package domain.models;

import org.deeplearning4j.nn.api.OptimizationAlgorithm;
import org.deeplearning4j.nn.conf.ComputationGraphConfiguration.GraphBuilder;
import org.deeplearning4j.nn.conf.ConvolutionMode;
import org.deeplearning4j.nn.conf.MultiLayerConfiguration;
import org.deeplearning4j.nn.conf.NeuralNetConfiguration;
import org.deeplearning4j.nn.conf.inputs.InputType;
import org.deeplearning4j.nn.conf.layers.Convolution3D;
import org.deeplearning4j.nn.conf.layers.ConvolutionLayer;

mckunkel

package faultrecordreader;

import static org.nd4j.linalg.indexing.NDArrayIndex.all;
import static org.nd4j.linalg.indexing.NDArrayIndex.point;

import java.io.DataInputStream;
import java.io.IOException;
import java.net.URI;
import java.util.ArrayList;
import java.util.Arrays;

mckunkel

public static ComputationGraph computationGraphModel(int numLabels) {

		GraphBuilder graphBuilder = new NeuralNetConfiguration.Builder().weightInit(WeightInit.XAVIER)
				.optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).updater(new Adam()).graphBuilder()
				.addInputs("input").setInputTypes(InputType.convolutionalFlat(112, 6, 1))
				.addLayer("cnn1",
						new ConvolutionLayer.Builder(3, 2).nIn(1).stride(1, 1).nOut(40).activation(new ActivationReLU())
								.build(),
						"input")
				.addLayer("cnn2",