Examples with GaussianDistribution org.nd4j.linalg.api.ops.random.impl.GaussianDistribution used on opensource projects

Example 1 with GaussianDistribution

use of org.nd4j.linalg.api.ops.random.impl.GaussianDistribution in project nd4j by deeplearning4j.

the class NormalDistribution method sample.

@Override
public INDArray sample(INDArray ret) {
    if (random.getStatePointer() != null) {
        if (means != null) {
            return Nd4j.getExecutioner().exec(new GaussianDistribution(ret, means, standardDeviation), random);
        } else {
            return Nd4j.getExecutioner().exec(new GaussianDistribution(ret, mean, standardDeviation), random);
        }
    } else {
        // For consistent values irrespective of c vs. fortran ordering
        Iterator<int[]> idxIter = new NdIndexIterator(ret.shape());
        int len = ret.length();
        if (means != null) {
            for (int i = 0; i < len; i++) {
                int[] idx = idxIter.next();
                ret.putScalar(idx, standardDeviation * random.nextGaussian() + means.getDouble(idx));
            }
        } else {
            for (int i = 0; i < len; i++) {
                ret.putScalar(idxIter.next(), standardDeviation * random.nextGaussian() + mean);
            }
        }
        return ret;
    }
}

Example 2 with GaussianDistribution

use of org.nd4j.linalg.api.ops.random.impl.GaussianDistribution in project nd4j by deeplearning4j.

the class NativeRandom method nextGaussian.

@Override
public INDArray nextGaussian(char order, int[] shape) {
    INDArray array = Nd4j.createUninitialized(shape, order);
    GaussianDistribution op = new GaussianDistribution(array, 0.0, 1.0);
    Nd4j.getExecutioner().exec(op, this);
    return array;
}

Example 3 with GaussianDistribution

use of org.nd4j.linalg.api.ops.random.impl.GaussianDistribution in project nd4j by deeplearning4j.

the class OrthogonalDistribution method sample.

@Override
public INDArray sample(int[] shape) {
    int numRows = 1;
    for (int i = 0; i < shape.length - 1; i++) numRows *= shape[i];
    int numCols = shape[shape.length - 1];
    val flatShape = new int[] { numRows, numCols };
    val flatRng = Nd4j.getExecutioner().exec(new GaussianDistribution(Nd4j.createUninitialized(flatShape, Nd4j.order()), 0.0, 1.0), random);
    int m = flatRng.rows();
    int n = flatRng.columns();
    val s = Nd4j.create(m < n ? m : n);
    val u = m < n ? Nd4j.create(m, n) : Nd4j.create(m, m);
    val v = Nd4j.create(n, n, 'f');
    Nd4j.getBlasWrapper().lapack().gesvd(flatRng, s, u, v);
    if (gains == null) {
        if (u.rows() == numRows && u.columns() == numCols) {
            return v.get(NDArrayIndex.interval(0, numRows), NDArrayIndex.interval(0, numCols)).mul(gain).reshape(shape);
        } else {
            return u.get(NDArrayIndex.interval(0, numRows), NDArrayIndex.interval(0, numCols)).mul(gain).reshape(shape);
        }
    } else {
        throw new UnsupportedOperationException();
    }
}

Example 4 with GaussianDistribution

use of org.nd4j.linalg.api.ops.random.impl.GaussianDistribution in project nd4j by deeplearning4j.

the class RandomProjection method gaussianRandomMatrix.

/**
 * Generate a dense Gaussian random matrix.
 *
 *   The n' components of the random matrix are drawn from
 *       N(0, 1.0 / n').
 *
 * @param shape
 * @param rng
 * @return
 */
private INDArray gaussianRandomMatrix(int[] shape, Random rng) {
    Nd4j.checkShapeValues(shape);
    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;
}