org.nd4j.linalg.api.rng.Random Java Exaples

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testTruncatedNormal1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z01 = Nd4j.create(10000000).assign(-119119d);
    INDArray z02 = Nd4j.createUninitialized(z01.length());

    TruncatedNormalDistribution distribution01 = new TruncatedNormalDistribution(z01, 0.0, 1.0);

    long time1 = System.currentTimeMillis();
    Nd4j.getExecutioner().exec(distribution01, random1);
    long time2 = System.currentTimeMillis();

    Nd4j.getExecutioner().exec(new GaussianDistribution( z02, 0.0, 1.0));
    long time3 = System.currentTimeMillis();

    log.info("Truncated: {} ms; Gaussian: {} ms", time2 - time1, time3 - time2);

    for (int e = 0; e < z01.length(); e++) {
        assertTrue("Value: " + z01.getDouble(e) + " at " + e,FastMath.abs(z01.getDouble(e)) <= 2.0);
        assertNotEquals(-119119d, z01.getDouble(e), 1e-3);
    }

    assertEquals(0.0, z01.meanNumber().doubleValue(), 1e-3);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testGaussianDistribution1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(100000);
    INDArray z2 = Nd4j.create(100000);
    INDArray zDup = z1.dup();

    GaussianDistribution op1 = new GaussianDistribution(z1, 0.0, 1.0);
    Nd4j.getExecutioner().exec(op1, random1);

    GaussianDistribution op2 = new GaussianDistribution(z2, 0.0, 1.0);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);
    assertEquals(0.0, z1.meanNumber().doubleValue(), 0.01);

    assertEquals(1.0, z1.stdNumber().doubleValue(), 0.01);

    assertEquals(z1, z2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testAlphaDropout1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.ones(300);
    INDArray z2 = Nd4j.ones(300);
    INDArray zDup = z1.dup();

    AlphaDropOut op1 = new AlphaDropOut(z1, z1, 0.10, 0.3, 0.5, 0.7);
    Nd4j.getExecutioner().exec(op1, random1);

    AlphaDropOut op2 = new AlphaDropOut(z2, z2, 0.10, 0.3, 0.5, 0.7);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);

    assertEquals(z1, z2);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testAlphaDropout1() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.ones(300);
    INDArray z2 = Nd4j.ones(300);
    INDArray zDup = z1.dup();

    AlphaDropOut op1 = new AlphaDropOut(z1, z1, 0.10, 0.3, 0.5, 0.7);
    Nd4j.getExecutioner().exec(op1, random1);

    AlphaDropOut op2 = new AlphaDropOut(z2, z2, 0.10, 0.3, 0.5, 0.7);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);

    assertEquals(z1, z2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testDropoutInverted1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.ones(300);
    INDArray z2 = Nd4j.ones(300);
    INDArray zDup = z1.dup();

    DropOutInverted op1 = new DropOutInverted(z1, z1, 0.10);
    Nd4j.getExecutioner().exec(op1, random1);

    DropOutInverted op2 = new DropOutInverted(z2, z2, 0.10);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);


    for (int x = 0; x < z1.length(); x++) {
        assertEquals("Failed on element: [" + x + "]", z1.getFloat(x), z2.getFloat(x), 0.01f);
    }
    assertEquals(z1, z2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testDistribution2() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(32);
    INDArray z2 = Nd4j.create(32);
    UniformDistribution distribution = new UniformDistribution(z1, 1.0, 2.0);
    Nd4j.getExecutioner().exec(distribution, random1);
    UniformDistribution distribution2 = new UniformDistribution(z2, 1.0, 2.0);
    Nd4j.getExecutioner().exec(distribution2, random2);

    System.out.println("Data: " + z1);
    System.out.println("Data: " + z2);
    for (int e = 0; e < z1.length(); e++) {
        double val = z1.getDouble(e);
        assertTrue(val >= 1.0 && val <= 2.0);
    }

    assertEquals(z1, z2);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testBinomialDistribution2() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(DataType.FLOAT, 1000);
    INDArray z2 = Nd4j.zeros(DataType.FLOAT,1000);
    INDArray z1Dup = Nd4j.zeros(DataType.FLOAT,1000);

    INDArray probs = Nd4j.create(new float[] {0.25f, 0.43f, 0.55f, 0.43f, 0.25f});

    BinomialDistribution op1 = new BinomialDistribution(z1, 5, probs);
    BinomialDistribution op2 = new BinomialDistribution(z2, 5, probs);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    BooleanIndexing.and(z1, Conditions.lessThanOrEqual(5.0));
    BooleanIndexing.and(z1, Conditions.greaterThanOrEqual(0.0));
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testGaussianDistribution3() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(DataType.DOUBLE, 1000000);
    INDArray z2 = Nd4j.create(DataType.DOUBLE, 1000000);

    GaussianDistribution op1 = new GaussianDistribution(z1, 1.0, 1.0);
    Nd4j.getExecutioner().exec(op1, random1);

    GaussianDistribution op2 = new GaussianDistribution(z2, -1.0, 2.0);
    Nd4j.getExecutioner().exec(op2, random2);


    assertEquals(1.0, z1.meanNumber().doubleValue(), 0.01);
    assertEquals(1.0, z1.stdNumber().doubleValue(), 0.01);

    // check variance
    assertEquals(-1.0, z2.meanNumber().doubleValue(), 0.01);
    assertEquals(4.0, z2.varNumber().doubleValue(), 0.01);

    assertNotEquals(z1, z2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testDistribution3() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(128);
    INDArray z2 = Nd4j.create(128);
    UniformDistribution distribution = new UniformDistribution(z1, 1.0, 2.0);
    Nd4j.getExecutioner().exec(distribution, random1);
    UniformDistribution distribution2 = new UniformDistribution(z2, 1.0, 2.0);
    Nd4j.getExecutioner().exec(distribution2, random1);

    System.out.println("Data: " + z1);
    System.out.println("Data: " + z2);

    assertNotEquals(z1, z2);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testBernoulliDistribution2() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.zeros(20);
    INDArray z2 = Nd4j.zeros(20);
    INDArray z1Dup = Nd4j.zeros(20);
    INDArray exp = Nd4j.create(new double[] {0,    1.0000,         0,    1.0000,    1.0000,         0,    1.0000,    1.0000,         0,    1.0000,    1.0000,    1.0000,         0,    1.0000,    1.0000,         0,         0,    1.0000,         0,    1.0000});

    BernoulliDistribution op1 = new BernoulliDistribution(z1, 0.50);
    BernoulliDistribution op2 = new BernoulliDistribution(z2, 0.50);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    assertEquals(exp, z1);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
    public void testDistribution1() {
        Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
        Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

        INDArray z1 = Nd4j.create(1000);
        INDArray z2 = Nd4j.create(1000);
        UniformDistribution distribution = new UniformDistribution(z1, 1.0, 2.0);
        Nd4j.getExecutioner().exec(distribution, random1);
        UniformDistribution distribution2 = new UniformDistribution(z2, 1.0, 2.0);
        Nd4j.getExecutioner().exec(distribution2, random2);

//        System.out.println("Data: " + z1);
//        System.out.println("Data: " + z2);
        for (int e = 0; e < z1.length(); e++) {
            double val = z1.getDouble(e);
            assertTrue(val >= 1.0 && val <= 2.0);
        }

        assertEquals(z1, z2);
    }

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testDropoutInverted1() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.ones(300);
    INDArray z2 = Nd4j.ones(300);
    INDArray zDup = z1.dup();

    DropOutInverted op1 = new DropOutInverted(z1, z1, 0.10);
    Nd4j.getExecutioner().exec(op1, random1);

    DropOutInverted op2 = new DropOutInverted(z2, z2, 0.10);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);


    for (int x = 0; x < z1.length(); x++) {
        assertEquals("Failed on element: [" + x + "]", z1.getFloat(x), z2.getFloat(x), 0.01f);
    }
    assertEquals(z1, z2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testBernoulliDistribution2() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.zeros(20);
    INDArray z2 = Nd4j.zeros(20);
    INDArray z1Dup = Nd4j.zeros(20);
    INDArray exp = Nd4j.create(new double[] {1.00, 0.00, 1.00, 0.00, 1.00, 1.00, 0.00, 1.00, 1.00, 0.00, 0.00, 0.00,
                    1.00, 1.00, 1.00, 0.00, 0.00, 0.00, 0.00, 0.00});

    BernoulliDistribution op1 = new BernoulliDistribution(z1, 0.50);
    BernoulliDistribution op2 = new BernoulliDistribution(z2, 0.50);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    assertEquals(exp, z1);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testBernoulliDistribution3() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray prob = Nd4j.create(new double[] {1.0, 0.1, 0.2, 0.5, 1.0, 1.0, 0.3, 0.7, 0.34, 0.119});

    INDArray z1 = Nd4j.zeros(10);
    INDArray z2 = Nd4j.zeros(10);
    INDArray z1Dup = Nd4j.zeros(10);
    INDArray exp = Nd4j.create(new double[] {1.00, 0.00, 1.00, 0.00, 1.00, 1.00, 0.00, 1.00, 1.00, 0.00});

    BernoulliDistribution op1 = new BernoulliDistribution(z1, prob);
    BernoulliDistribution op2 = new BernoulliDistribution(z2, prob);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    assertEquals(exp, z1);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testBinomialDistribution1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.zeros(1000);
    INDArray z2 = Nd4j.zeros(1000);
    INDArray z1Dup = Nd4j.zeros(1000);

    BinomialDistribution op1 = new BinomialDistribution(z1, 5, 0.25);
    BinomialDistribution op2 = new BinomialDistribution(z2, 5, 0.25);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    BooleanIndexing.and(z1, Conditions.lessThanOrEqual(5.0));
    BooleanIndexing.and(z1, Conditions.greaterThanOrEqual(0.0));
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testBernoulliDistribution3() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray prob = Nd4j.create(new double[] {1.0, 0.1, 0.2, 0.5, 1.0, 1.0, 0.3, 0.7, 0.34, 0.119});

    INDArray z1 = Nd4j.zeros(10);
    INDArray z2 = Nd4j.zeros(10);
    INDArray z1Dup = Nd4j.zeros(10);
    INDArray exp = Nd4j.create(new double[] {1.0000,         0,         0,    1.0000,    1.0000,    1.0000,         0,    1.0000,         0,         0});

    BernoulliDistribution op1 = new BernoulliDistribution(z1, prob);
    BernoulliDistribution op2 = new BernoulliDistribution(z2, prob);

    Nd4j.getExecutioner().exec(op1, random1);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(z1Dup, z1);

    assertEquals(z1, z2);

    assertEquals(exp, z1);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
@Ignore
public void testTruncatedNormal1() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z01 = Nd4j.create(10000000).assign(-119119d);
    INDArray z02 = Nd4j.createUninitialized(z01.length());

    TruncatedNormalDistribution distribution01 = new TruncatedNormalDistribution(z01, 0.0, 1.0);

    long time1 = System.currentTimeMillis();
    Nd4j.getExecutioner().exec(distribution01, random1);
    long time2 = System.currentTimeMillis();

    Nd4j.getExecutioner().exec(new GaussianDistribution( z02, 0.0, 1.0));
    long time3 = System.currentTimeMillis();

    log.info("Truncated: {} ms; Gaussian: {} ms", time2 - time1, time3 - time2);

    for (int e = 0; e < z01.length(); e++) {
        assertTrue("Value: " + z01.getDouble(e) + " at " + e,FastMath.abs(z01.getDouble(e)) <= 2.0);
        assertNotEquals(-119119d, z01.getDouble(e), 1e-3);
    }

    assertEquals(0.0, z01.meanNumber().doubleValue(), 1e-3);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

6 votes

@Test
public void testJavaSide1() throws Exception {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    float array1[] = new float[1000];
    float array2[] = new float[1000];

    for (int e = 0; e < array1.length; e++) {
        array1[e] = random1.nextFloat();
        array2[e] = random2.nextFloat();

        assertTrue(array1[e] <= 1.0f);
    }

    assertArrayEquals(array1, array2, 1e-5f);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
    public void testDistribution3() {
        Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);

        INDArray z1 = Nd4j.create(128);
        INDArray z2 = Nd4j.create(128);
        UniformDistribution distribution = new UniformDistribution(z1, 1.0, 2.0);
        Nd4j.getExecutioner().exec(distribution, random1);
        UniformDistribution distribution2 = new UniformDistribution(z2, 1.0, 2.0);
        Nd4j.getExecutioner().exec(distribution2, random1);

//        System.out.println("Data: " + z1);
//        System.out.println("Data: " + z2);

        assertNotEquals(z1, z2);
    }

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

6 votes

@Test
public void testJavaSide2() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    int array1[] = new int[1000];
    int array2[] = new int[1000];

    for (int e = 0; e < array1.length; e++) {
        array1[e] = random1.nextInt();
        array2[e] = random2.nextInt();

        assertEquals(array1[e], array2[e]);
        assertTrue(array1[e] >= 0);
    }

    assertArrayEquals(array1, array2);
}

Source File: RandomTests.java From nd4j with Apache License 2.0

5 votes

@Test
public void testMultithreading1() throws Exception {

    final AtomicInteger cnt = new AtomicInteger(0);
    final CopyOnWriteArrayList<float[]> list = new CopyOnWriteArrayList<>();

    Thread[] threads = new Thread[10];
    for (int x = 0; x < threads.length; x++) {
        list.add(null);
    }

    for (int x = 0; x < threads.length; x++) {
        threads[x] = new Thread(new Runnable() {
            @Override
            public void run() {
                Random rnd = Nd4j.getRandom();
                rnd.setSeed(119);
                float[] array = new float[10];

                for (int e = 0; e < array.length; e++) {
                    array[e] = rnd.nextFloat();
                }
                list.set(cnt.getAndIncrement(), array);
            }
        });
        threads[x].start();
    }

    for (int x = 0; x < threads.length; x++) {
        threads[x].join();

        assertNotEquals(null, list.get(x));

        if (x > 0) {
            assertArrayEquals(list.get(0), list.get(x), 1e-5f);
        }
    }
}

Source File: InMemoryLookupTable.java From deeplearning4j with Apache License 2.0

5 votes

public InMemoryLookupTable(VocabCache<T> vocab, int vectorLength, boolean useAdaGrad, double lr, Random gen,
                double negative) {
    this.vocab = vocab;
    this.vectorLength = vectorLength;
    this.useAdaGrad = useAdaGrad;
    this.lr.set(lr);
    this.rng = gen;
    this.negative = negative;
    initExpTable();

    if (useAdaGrad) {
        initAdaGrad();
    }
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

5 votes

@Test
public void testMultithreading2() throws Exception {

    final AtomicInteger cnt = new AtomicInteger(0);
    final CopyOnWriteArrayList<INDArray> list = new CopyOnWriteArrayList<>();

    Thread[] threads = new Thread[10];
    for (int x = 0; x < threads.length; x++) {
        list.add(null);
    }

    for (int x = 0; x < threads.length; x++) {
        threads[x] = new Thread(new Runnable() {
            @Override
            public void run() {
                Random rnd = Nd4j.getRandom();
                rnd.setSeed(119);
                INDArray array = Nd4j.getExecutioner().exec(new UniformDistribution(Nd4j.createUninitialized(25)));

                Nd4j.getExecutioner().commit();

                list.set(cnt.getAndIncrement(), array);
            }
        });
        threads[x].start();
    }

    // we want all threads finished before comparing arrays
    for (int x = 0; x < threads.length; x++)
        threads[x].join();

    for (int x = 0; x < threads.length; x++) {
        assertNotEquals(null, list.get(x));

        if (x > 0) {
            assertEquals(list.get(0), list.get(x));
        }
    }
}

Source File: QLearningDiscrete.java From deeplearning4j with Apache License 2.0

5 votes

public QLearningDiscrete(MDP<O, Integer, DiscreteSpace> mdp, IDQN dqn, QLearningConfiguration conf,
                         int epsilonNbStep, ILearningBehavior<Integer> learningBehavior, Random random) {
    this.configuration = conf;
    this.mdp = new LegacyMDPWrapper<>(mdp, null);
    qNetwork = dqn;
    policy = new DQNPolicy(getQNetwork());
    egPolicy = new EpsGreedy(policy, mdp, conf.getUpdateStart(), epsilonNbStep, random, conf.getMinEpsilon(),
            this);

    this.learningBehavior = learningBehavior;
}

Source File: Nd4jTest.java From nd4j with Apache License 2.0

5 votes

@Test
public void testGetRandomSetSeed() {
    Random r = Nd4j.getRandom();
    Random t = Nd4j.getRandom();

    assertEquals(r, t);
    r.setSeed(123);
    assertEquals(r, t);
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

5 votes

@Test
public void testGaussianDistribution1() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    INDArray z1 = Nd4j.create(DataType.DOUBLE, 1000000);
    INDArray z2 = Nd4j.create(DataType.DOUBLE, 1000000);
    INDArray zDup = z1.like();

    GaussianDistribution op1 = new GaussianDistribution(z1, 0.0, 1.0);
    Nd4j.getExecutioner().exec(op1, random1);

    GaussianDistribution op2 = new GaussianDistribution(z2, 0.0, 1.0);
    Nd4j.getExecutioner().exec(op2, random2);

    assertNotEquals(zDup, z1);
    assertEquals(0.0, z1.meanNumber().doubleValue(), 0.01);

    assertEquals(1.0, z1.stdNumber().doubleValue(), 0.01);

    val d1 = z1.toDoubleVector();
    val d2 = z2.toDoubleVector();

    assertArrayEquals(d1, d2, 1e-4);

    assertEquals(z1, z2);
}

Source File: NormalDistribution.java From nd4j with Apache License 2.0

5 votes

/**
 * Creates a normal distribution.
 *
 * @param rng                Random number generator.
 * @param mean               Mean for this distribution.
 * @param sd                 Standard deviation for this distribution.
 * @param inverseCumAccuracy Inverse cumulative probability accuracy.
 * @throws NotStrictlyPositiveException if {@code sd <= 0}.
 * @since 3.1
 */
public NormalDistribution(Random rng, double mean, double sd, double inverseCumAccuracy)
                throws NotStrictlyPositiveException {
    super(rng);

    if (sd <= 0) {
        throw new NotStrictlyPositiveException(LocalizedFormats.STANDARD_DEVIATION, sd);
    }

    this.mean = mean;
    standardDeviation = sd;
    solverAbsoluteAccuracy = inverseCumAccuracy;
}

Source File: TruncatedNormalDistribution.java From nd4j with Apache License 2.0

5 votes

/**
 * Creates a normal distribution.
 *
 * @param rng                Random number generator.
 * @param mean               Mean for this distribution.
 * @param sd                 Standard deviation for this distribution.
 * @param inverseCumAccuracy Inverse cumulative probability accuracy.
 * @throws NotStrictlyPositiveException if {@code sd <= 0}.
 * @since 3.1
 */
public TruncatedNormalDistribution(Random rng, double mean, double sd, double inverseCumAccuracy)
                throws NotStrictlyPositiveException {
    super(rng);

    if (sd <= 0) {
        throw new NotStrictlyPositiveException(LocalizedFormats.STANDARD_DEVIATION, sd);
    }

    this.mean = mean;
    standardDeviation = sd;
    solverAbsoluteAccuracy = inverseCumAccuracy;
}

Source File: RandomTests.java From deeplearning4j with Apache License 2.0

5 votes

/**
 * This test checks reSeed mechanics for native side
 *
 * @throws Exception
 */

@Test
public void testJavaSide4() {
    Random random1 = Nd4j.getRandomFactory().getNewRandomInstance(119);
    Random random2 = Nd4j.getRandomFactory().getNewRandomInstance(119);

    int array1[] = new int[1000];
    int array2[] = new int[1000];

    for (int e = 0; e < array1.length; e++) {
        array1[e] = random1.nextInt();
        array2[e] = random2.nextInt();

        assertEquals(array1[e], array2[e]);
        assertTrue(array1[e] >= 0);
    }

    assertArrayEquals(array1, array2);

    random1.reSeed();
    random1.reSeed();

    int array3[] = new int[1000];
    int array4[] = new int[1000];

    for (int e = 0; e < array1.length; e++) {
        array3[e] = random1.nextInt();
        array4[e] = random2.nextInt();

        assertNotEquals(array3[e], array4[e]);
        assertTrue(array1[e] >= 0);
    }
}

Source File: RandomProjectionLSH.java From deeplearning4j with Apache License 2.0

5 votes

private INDArray gaussianRandomMatrix(int[] shape, Random rng){
    INDArray res = Nd4j.create(shape);

    GaussianDistribution op1 = new GaussianDistribution(res, 0.0, 1.0 / Math.sqrt(shape[0]));

    Nd4j.getExecutioner().exec(op1, rng);
    return res;
}

org.nd4j.linalg.api.rng.Random Java Examples