gs-cv: Test a saved network

gs-cv/pom.xml

@@ -260,17 +260,11 @@
 			<groupId>org.deeplearning4j</groupId>
 			<artifactId>deeplearning4j-core</artifactId>
 			<version>${deeplearning4j.version}</version>
-			<exclusions>
-				<exclusion>
-					<groupId>org.bytedeco.javacpp-presets</groupId>
-					<artifactId>opencv-platform</artifactId>
-				</exclusion>
-			</exclusions>
 		</dependency>
 
 		<dependency>
 			<groupId>org.nd4j</groupId>
-			<artifactId>nd4j-cuda-8.0-platform</artifactId>
+			<artifactId>nd4j-cuda-8.0</artifactId>
 			<version>${deeplearning4j.version}</version>
 		</dependency>
 

gs-cv/src/main/java/org/genericsystem/cv/nn/TestNet.java

@@ -0,0 +1,71 @@
+package org.genericsystem.cv.nn;
+
+import java.io.File;
+import java.io.IOException;
+import java.util.List;
+import java.util.Random;
+
+import org.datavec.api.io.filters.BalancedPathFilter;
+import org.datavec.api.io.labels.ParentPathLabelGenerator;
+import org.datavec.api.split.FileSplit;
+import org.datavec.api.split.InputSplit;
+import org.datavec.image.loader.BaseImageLoader;
+import org.datavec.image.loader.NativeImageLoader;
+import org.datavec.image.recordreader.ImageRecordReader;
+import org.deeplearning4j.nn.graph.ComputationGraph;
+import org.deeplearning4j.util.ModelSerializer;
+import org.genericsystem.cv.utils.NativeLibraryLoader;
+import org.nd4j.linalg.api.ndarray.INDArray;
+import org.nd4j.linalg.dataset.api.preprocessor.DataNormalization;
+import org.nd4j.linalg.dataset.api.preprocessor.ImagePreProcessingScaler;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+public class TestNet {
+	private static final Logger log = LoggerFactory.getLogger(TestNet.class);
+
+	static {
+		NativeLibraryLoader.load();
+	}
+
+	// Predict result for an image.
+	// Problem: Output array does not include labels.
+	public static void classifyImage(File imageFile, List<String> labels) {
+		ComputationGraph graph = getComputationGraph(new File("AdaptedVGG16/AdaptedVGG16-grouped-acc-1.zip"));
+
+		NativeImageLoader loader = new NativeImageLoader(224, 224, 3);
+		DataNormalization scaler = new ImagePreProcessingScaler(-1, 1);
+		try {
+			INDArray image = loader.asMatrix(imageFile);
+			scaler.transform(image);
+			log.info("Result: {}.", graph.outputSingle(image));
+		} catch (IOException e) {
+			throw new RuntimeException("Impossible to load image " + imageFile, e);
+		}
+	}
+
+	public static void main(String[] args) {
+		File parentDir = new File(System.getProperty("user.dir"), "data/training-grouped-augmented");
+		FileSplit filesInDir = new FileSplit(parentDir, BaseImageLoader.ALLOWED_FORMATS);
+		ParentPathLabelGenerator labelMaker = new ParentPathLabelGenerator();
+		BalancedPathFilter pathFilter = new BalancedPathFilter(new Random(123), BaseImageLoader.ALLOWED_FORMATS, labelMaker, 0, 0, 100, 0);
+
+		InputSplit[] filesInDirSplit = filesInDir.sample(pathFilter);
+		InputSplit testData = filesInDirSplit[0];
+		try (ImageRecordReader recordReader = new ImageRecordReader(224, 224, 3, labelMaker)) {
+			recordReader.initialize(testData, null);
+			classifyImage(new File("data/validation/id-fr-front/dimage-1.png"), recordReader.getLabels());
+		} catch (IOException e) {
+			log.error("Impossible to load data", e);
+			return;
+		}
+	}
+
+	public static ComputationGraph getComputationGraph(File modelFile) {
+		try {
+			return ModelSerializer.restoreComputationGraph(modelFile);
+		} catch (IOException e) {
+			throw new RuntimeException("Impossible to load model from disk.", e);
+		}
+	}
+}