Java NormalizerStandardize类代码示例

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
private void createDataSource() throws IOException, InterruptedException {
    //First: get the dataset using the record reader. CSVRecordReader handles loading/parsing
    int numLinesToSkip = 0;
    String delimiter = ",";
    RecordReader recordReader = new CSVRecordReader(numLinesToSkip, delimiter);
    recordReader.initialize(new InputStreamInputSplit(dataFile));

    //Second: the RecordReaderDataSetIterator handles conversion to DataSet objects, ready for use in neural network
    int labelIndex = 4;     //5 values in each row of the iris.txt CSV: 4 input features followed by an integer label (class) index. Labels are the 5th value (index 4) in each row
    int numClasses = 3;     //3 classes (types of iris flowers) in the iris data set. Classes have integer values 0, 1 or 2

    DataSetIterator iterator = new RecordReaderDataSetIterator(recordReader, batchSize, labelIndex, numClasses);
    DataSet allData = iterator.next();
    allData.shuffle();

    SplitTestAndTrain testAndTrain = allData.splitTestAndTrain(0.80);  //Use 80% of data for training

    trainingData = testAndTrain.getTrain();
    testData = testAndTrain.getTest();

    //We need to normalize our data. We'll use NormalizeStandardize (which gives us mean 0, unit variance):
    DataNormalization normalizer = new NormalizerStandardize();
    normalizer.fit(trainingData);           //Collect the statistics (mean/stdev) from the training data. This does not modify the input data
    normalizer.transform(trainingData);     //Apply normalization to the training data
    normalizer.transform(testData);         //Apply normalization to the test data. This is using statistics calculated from the *training* set
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
private void createDataSource() throws IOException, InterruptedException {
    //First: get the dataset using the record reader. CSVRecordReader handles loading/parsing
    int numLinesToSkip = 0;
    String delimiter = ",";
    RecordReader recordReader = new CSVRecordReader(numLinesToSkip, delimiter);
    recordReader.initialize(new InputStreamInputSplit(dataFile));

    //Second: the RecordReaderDataSetIterator handles conversion to DataSet objects, ready for use in neural network
    int labelIndex = 11;

    DataSetIterator iterator = new RecordReaderDataSetIterator(recordReader, batchSize, labelIndex, labelIndex, true);
    DataSet allData = iterator.next();

    SplitTestAndTrain testAndTrain = allData.splitTestAndTrain(0.80);  //Use 80% of data for training

    trainingData = testAndTrain.getTrain();
    testData = testAndTrain.getTest();

    //We need to normalize our data. We'll use NormalizeStandardize (which gives us mean 0, unit variance):
    DataNormalization normalizer = new NormalizerStandardize();
    normalizer.fit(trainingData);           //Collect the statistics (mean/stdev) from the training data. This does not modify the input data
    normalizer.transform(trainingData);     //Apply normalization to the training data
    normalizer.transform(testData);         //Apply normalization to the test data. This is using statistics calculated from the *training* set
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testBruteForce4d() {
    Construct4dDataSet imageDataSet = new Construct4dDataSet(10, 5, 10, 15);

    NormalizerStandardize myNormalizer = new NormalizerStandardize();
    myNormalizer.fit(imageDataSet.sampleDataSet);
    assertEquals(imageDataSet.expectedMean, myNormalizer.getMean());

    float aat = Transforms.abs(myNormalizer.getStd().div(imageDataSet.expectedStd).sub(1)).maxNumber().floatValue();
    float abt = myNormalizer.getStd().maxNumber().floatValue();
    float act = imageDataSet.expectedStd.maxNumber().floatValue();
    System.out.println("ValA: " + aat);
    System.out.println("ValB: " + abt);
    System.out.println("ValC: " + act);
    assertTrue(aat < 0.05);

    NormalizerMinMaxScaler myMinMaxScaler = new NormalizerMinMaxScaler();
    myMinMaxScaler.fit(imageDataSet.sampleDataSet);
    assertEquals(imageDataSet.expectedMin, myMinMaxScaler.getMin());
    assertEquals(imageDataSet.expectedMax, myMinMaxScaler.getMax());

    DataSet copyDataSet = imageDataSet.sampleDataSet.copy();
    myNormalizer.transform(copyDataSet);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testRevert() {
    double tolerancePerc = 0.01; // 0.01% of correct value
    int nSamples = 500;
    int nFeatures = 3;

    INDArray featureSet = Nd4j.randn(nSamples, nFeatures);
    INDArray labelSet = Nd4j.zeros(nSamples, 1);
    DataSet sampleDataSet = new DataSet(featureSet, labelSet);

    NormalizerStandardize myNormalizer = new NormalizerStandardize();
    myNormalizer.fit(sampleDataSet);
    DataSet transformed = sampleDataSet.copy();
    myNormalizer.transform(transformed);
    //System.out.println(transformed.getFeatures());
    myNormalizer.revert(transformed);
    //System.out.println(transformed.getFeatures());
    INDArray delta = Transforms.abs(transformed.getFeatures().sub(sampleDataSet.getFeatures()))
                    .div(sampleDataSet.getFeatures());
    double maxdeltaPerc = delta.max(0, 1).mul(100).getDouble(0, 0);
    assertTrue(maxdeltaPerc < tolerancePerc);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testRestoreUnsavedNormalizerFromInputStream() throws Exception {
    DataSet dataSet = trivialDataSet();

    NormalizerStandardize norm = new NormalizerStandardize();
    norm.fit(dataSet);

    ComputationGraph cg = simpleComputationGraph();
    cg.init();

    File tempFile = File.createTempFile("tsfs", "fdfsdf");
    tempFile.deleteOnExit();
    ModelSerializer.writeModel(cg, tempFile, true);

    FileInputStream fis = new FileInputStream(tempFile);

    NormalizerStandardize restored = ModelSerializer.restoreNormalizerFromInputStream(fis);

    assertEquals(null, restored);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testMeanStdZeros() {
    List<List<Writable>> data = new ArrayList<>();
    Schema.Builder builder = new Schema.Builder();
    int numColumns = 6;
    for (int i = 0; i < numColumns; i++)
        builder.addColumnDouble(String.valueOf(i));

    for (int i = 0; i < 5; i++) {
        List<Writable> record = new ArrayList<>(numColumns);
        data.add(record);
        for (int j = 0; j < numColumns; j++) {
            record.add(new DoubleWritable(1.0));
        }

    }

    INDArray arr = RecordConverter.toMatrix(data);

    Schema schema = builder.build();
    JavaRDD<List<Writable>> rdd = sc.parallelize(data);
    DataRowsFacade dataFrame = DataFrames.toDataFrame(schema, rdd);

    //assert equivalent to the ndarray pre processing
    NormalizerStandardize standardScaler = new NormalizerStandardize();
    standardScaler.fit(new DataSet(arr.dup(), arr.dup()));
    INDArray standardScalered = arr.dup();
    standardScaler.transform(new DataSet(standardScalered, standardScalered));
    DataNormalization zeroToOne = new NormalizerMinMaxScaler();
    zeroToOne.fit(new DataSet(arr.dup(), arr.dup()));
    INDArray zeroToOnes = arr.dup();
    zeroToOne.transform(new DataSet(zeroToOnes, zeroToOnes));
    List<Row> rows = Normalization.stdDevMeanColumns(dataFrame, dataFrame.get().columns());
    INDArray assertion = DataFrames.toMatrix(rows);
    //compare standard deviation
    assertTrue(standardScaler.getStd().equalsWithEps(assertion.getRow(0), 1e-1));
    //compare mean
    assertTrue(standardScaler.getMean().equalsWithEps(assertion.getRow(1), 1e-1));

}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
static DataIterator<NormalizerStandardize> irisCsv(String name) {
    CSVRecordReader recordReader = new CSVRecordReader(0, ",");
    try {
        recordReader.initialize(new FileSplit(new File(name)));
    } catch (Exception e) {
        e.printStackTrace();
    }

    int labelIndex = 4;     //5 values in each row of the iris.txt CSV: 4 input features followed by an integer label (class) index. Labels are the 5th value (index 4) in each row
    int numClasses = 3;     //3 classes (types of iris flowers) in the iris data set. Classes have integer values 0, 1 or 2
    int batchSize = 50;     //Iris data set: 150 examples total.

    RecordReaderDataSetIterator iterator = new RecordReaderDataSetIterator(
            recordReader,
            batchSize,
            labelIndex,
            numClasses
    );

    NormalizerStandardize normalizer = new NormalizerStandardize();

    while (iterator.hasNext()) {
        normalizer.fit(iterator.next());
    }
    iterator.reset();

    iterator.setPreProcessor(normalizer);

    return new DataIterator<>(iterator, normalizer);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
public static void main(String... args) throws Exception {
    Options options = new Options();

    options.addOption("i", "input", true, "The file with test data.");
    options.addOption("m", "model", true, "Name of trained model file.");

    CommandLine cmd = new BasicParser().parse(options, args);

    String input = cmd.getOptionValue("i");
    String modelName = cmd.getOptionValue("m");

    if (cmd.hasOption("i") && cmd.hasOption("m")) {
        MultiLayerNetwork model = ModelSerializer.restoreMultiLayerNetwork(modelName);
        DataIterator<NormalizerStandardize> it = DataIterator.irisCsv(input);
        RecordReaderDataSetIterator testData = it.getIterator();
        NormalizerStandardize normalizer = it.getNormalizer();
        normalizer.load(
                new File(modelName + ".norm1"),
                new File(modelName + ".norm2"),
                new File(modelName + ".norm3"),
                new File(modelName + ".norm4")
        );

        Evaluation eval = new Evaluation(3);
        while (testData.hasNext()) {
            DataSet ds = testData.next();
            INDArray output = model.output(ds.getFeatureMatrix());
            eval.eval(ds.getLabels(), output);
        }

        log.info(eval.stats());
    } else {
        log.error("Invalid arguments.");

        new HelpFormatter().printHelp("Evaluate", options);
    }
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testBruteForce3d() {

    NormalizerStandardize myNormalizer = new NormalizerStandardize();
    NormalizerMinMaxScaler myMinMaxScaler = new NormalizerMinMaxScaler();

    int timeSteps = 15;
    int samples = 100;
    //multiplier for the features
    INDArray featureScaleA = Nd4j.create(new double[] {1, -2, 3}).reshape(3, 1);
    INDArray featureScaleB = Nd4j.create(new double[] {2, 2, 3}).reshape(3, 1);

    Construct3dDataSet caseA = new Construct3dDataSet(featureScaleA, timeSteps, samples, 1);
    Construct3dDataSet caseB = new Construct3dDataSet(featureScaleB, timeSteps, samples, 1);

    myNormalizer.fit(caseA.sampleDataSet);
    assertEquals(caseA.expectedMean, myNormalizer.getMean());
    assertTrue(Transforms.abs(myNormalizer.getStd().div(caseA.expectedStd).sub(1)).maxNumber().floatValue() < 0.01);

    myMinMaxScaler.fit(caseB.sampleDataSet);
    assertEquals(caseB.expectedMin, myMinMaxScaler.getMin());
    assertEquals(caseB.expectedMax, myMinMaxScaler.getMax());

    //Same Test with an Iterator, values should be close for std, exact for everything else
    DataSetIterator sampleIterA = new TestDataSetIterator(caseA.sampleDataSet, 5);
    DataSetIterator sampleIterB = new TestDataSetIterator(caseB.sampleDataSet, 5);

    myNormalizer.fit(sampleIterA);
    assertEquals(myNormalizer.getMean(), caseA.expectedMean);
    assertTrue(Transforms.abs(myNormalizer.getStd().div(caseA.expectedStd).sub(1)).maxNumber().floatValue() < 0.01);

    myMinMaxScaler.fit(sampleIterB);
    assertEquals(myMinMaxScaler.getMin(), caseB.expectedMin);
    assertEquals(myMinMaxScaler.getMax(), caseB.expectedMax);

}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testDifferentBatchSizes() {
    // Create 6x1 matrix of the numbers 1 through 6
    INDArray values = Nd4j.linspace(1, 6, 6).transpose();
    DataSet dataSet = new DataSet(values, values);

    // Test fitting a DataSet
    NormalizerStandardize norm1 = new NormalizerStandardize();
    norm1.fit(dataSet);
    assertEquals(3.5f, norm1.getMean().getFloat(0), 1e-6);
    assertEquals(1.70783f, norm1.getStd().getFloat(0), 1e-4);

    // Test fitting an iterator with equal batch sizes
    DataSetIterator testIter1 = new TestDataSetIterator(dataSet, 3); // Will yield 2 batches of 3 rows
    NormalizerStandardize norm2 = new NormalizerStandardize();
    norm2.fit(testIter1);
    assertEquals(3.5f, norm2.getMean().getFloat(0), 1e-6);
    assertEquals(1.70783f, norm2.getStd().getFloat(0), 1e-4);

    // Test fitting an iterator with varying batch sizes
    DataSetIterator testIter2 = new TestDataSetIterator(dataSet, 4); // Will yield batch of 4 and batch of 2 rows
    NormalizerStandardize norm3 = new NormalizerStandardize();
    norm3.fit(testIter2);
    assertEquals(3.5f, norm3.getMean().getFloat(0), 1e-6);
    assertEquals(1.70783f, norm3.getStd().getFloat(0), 1e-4);

    // Test fitting an iterator with batches of single rows
    DataSetIterator testIter3 = new TestDataSetIterator(dataSet, 1); // Will yield 6 batches of 1 row
    NormalizerStandardize norm4 = new NormalizerStandardize();
    norm4.fit(testIter3);
    assertEquals(3.5f, norm4.getMean().getFloat(0), 1e-6);
    assertEquals(1.70783f, norm4.getStd().getFloat(0), 1e-4);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testUnderOverflow() {
    // This dataset will be basically constant with a small std deviation
    // And the constant is large. Checking if algorithm can handle
    double tolerancePerc = 1; //Within 1 %
    double toleranceAbs = 0.0005;
    int nSamples = 1000;
    int bSize = 10;
    int x = -1000000, y = 1000000;
    double z = 1000000;

    INDArray featureX = Nd4j.rand(nSamples, 1).mul(1).add(x);
    INDArray featureY = Nd4j.rand(nSamples, 1).mul(2).add(y);
    INDArray featureZ = Nd4j.rand(nSamples, 1).mul(3).add(z);
    INDArray featureSet = Nd4j.concat(1, featureX, featureY, featureZ);
    INDArray labelSet = Nd4j.zeros(nSamples, 1);
    DataSet sampleDataSet = new DataSet(featureSet, labelSet);
    DataSetIterator sampleIter = new TestDataSetIterator(sampleDataSet, bSize);

    INDArray theoreticalMean = Nd4j.create(new double[] {x, y, z});

    NormalizerStandardize myNormalizer = new NormalizerStandardize();
    myNormalizer.fit(sampleIter);

    INDArray meanDelta = Transforms.abs(theoreticalMean.sub(myNormalizer.getMean()));
    INDArray meanDeltaPerc = meanDelta.mul(100).div(theoreticalMean);
    assertTrue(meanDeltaPerc.max(1).getDouble(0, 0) < tolerancePerc);

    //this just has to not barf
    //myNormalizer.transform(sampleIter);
    myNormalizer.transform(sampleDataSet);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Test
public void testConstant() {
    double tolerancePerc = 10.0; // 10% of correct value
    int nSamples = 500;
    int nFeatures = 3;
    int constant = 100;

    INDArray featureSet = Nd4j.zeros(nSamples, nFeatures).add(constant);
    INDArray labelSet = Nd4j.zeros(nSamples, 1);
    DataSet sampleDataSet = new DataSet(featureSet, labelSet);


    NormalizerStandardize myNormalizer = new NormalizerStandardize();
    myNormalizer.fit(sampleDataSet);
    //Checking if we gets nans
    assertFalse(Double.isNaN(myNormalizer.getStd().getDouble(0)));

    myNormalizer.transform(sampleDataSet);
    //Checking if we gets nans, because std dev is zero
    assertFalse(Double.isNaN(sampleDataSet.getFeatures().min(0, 1).getDouble(0)));
    //Checking to see if transformed values are close enough to zero
    assertEquals(Transforms.abs(sampleDataSet.getFeatures()).max(0, 1).getDouble(0, 0), 0,
                    constant * tolerancePerc / 100.0);

    myNormalizer.revert(sampleDataSet);
    //Checking if we gets nans, because std dev is zero
    assertFalse(Double.isNaN(sampleDataSet.getFeatures().min(0, 1).getDouble(0)));
    assertEquals(Transforms.abs(sampleDataSet.getFeatures().sub(featureSet)).min(0, 1).getDouble(0), 0,
                    constant * tolerancePerc / 100.0);
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Before
public void randomData() {
    Nd4j.getRandom().setSeed(12345);
    batchSize = 13;
    batchCount = 20;
    lastBatch = batchSize / 2;
    INDArray origFeatures = Nd4j.rand(batchCount * batchSize + lastBatch, 10);
    INDArray origLabels = Nd4j.rand(batchCount * batchSize + lastBatch, 3);
    data = new DataSet(origFeatures, origLabels);
    stdScaler = new NormalizerStandardize();
    minMaxScaler = new NormalizerMinMaxScaler();
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Override
public void write(@NonNull NormalizerStandardize normalizer, @NonNull OutputStream stream) throws IOException {
    try (DataOutputStream dos = new DataOutputStream(stream)) {
        dos.writeBoolean(normalizer.isFitLabel());

        Nd4j.write(normalizer.getMean(), dos);
        Nd4j.write(normalizer.getStd(), dos);

        if (normalizer.isFitLabel()) {
            Nd4j.write(normalizer.getLabelMean(), dos);
            Nd4j.write(normalizer.getLabelStd(), dos);
        }
        dos.flush();
    }
}
 

import org.nd4j.linalg.dataset.api.preprocessor.NormalizerStandardize; //导入依赖的package包/类
@Override
public NormalizerStandardize restore(@NonNull InputStream stream) throws IOException {
    DataInputStream dis = new DataInputStream(stream);

    boolean fitLabels = dis.readBoolean();

    NormalizerStandardize result = new NormalizerStandardize(Nd4j.read(dis), Nd4j.read(dis));
    result.fitLabel(fitLabels);
    if (fitLabels) {
        result.setLabelStats(Nd4j.read(dis), Nd4j.read(dis));
    }

    return result;
}