org.nd4j.linalg.factory.Nd4j.getDistributions()方法的使用及代码示例

x33g5p2x  于2022-01-24 转载在 其他  
字(7.6k)|赞(0)|评价(0)|浏览(91)

本文整理了Java中org.nd4j.linalg.factory.Nd4j.getDistributions()方法的一些代码示例，展示了Nd4j.getDistributions()的具体用法。这些代码示例主要来源于Github/Stackoverflow/Maven等平台，是从一些精选项目中提取出来的代码，具有较强的参考意义，能在一定程度帮忙到你。Nd4j.getDistributions()方法的具体详情如下：
包路径：org.nd4j.linalg.factory.Nd4j
类名称：Nd4j
方法名：getDistributions

Nd4j.getDistributions介绍

[英]Get the primary distributions factory
[中]获取主分发工厂

代码示例

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  //As per Glorot and Bengio 2010: Uniform distribution U(-s,s) with s = sqrt(6/(fanIn + fanOut))
  //Eq 16: http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf
  double s = Math.sqrt(6.0) / Math.sqrt(fanIn + fanOut);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-s, s));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double a = 1.0 / Math.sqrt(fanIn);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-a, a));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double r = 4.0 * Math.sqrt(6.0 / (fanIn + fanOut));
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-r, r));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double b = 3.0 / Math.sqrt(fanIn);
  return  Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-b, b));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double scalingFanIn = 3.0 / Math.sqrt(fanIn);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-scalingFanIn, scalingFanIn));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double u = Math.sqrt(6.0 / fanIn);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-u, u)); //U(-sqrt(6/fanIn), sqrt(6/fanIn)
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double scalingFanAvg = 3.0 / Math.sqrt((fanIn + fanOut) / 2);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-scalingFanAvg, scalingFanAvg));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray doCreate(long[] shape, INDArray paramsView) {
  double scalingFanOut = 3.0 / Math.sqrt(fanOut);
  return Nd4j.rand(shape, Nd4j.getDistributions().createUniform(-scalingFanOut, scalingFanOut));
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray rand(long[] shape, float min, float max, org.nd4j.linalg.api.rng.Random rng) {
  //ensure shapes that wind up being scalar end up with the write shape
  if (shape.length == 1 && shape[0] == 0) {
    shape = new long[] {1, 1};
  }
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: deeplearning4j/nd4j

/**
 * Generates a random matrix between min and max
 *
 * @param shape the number of rows of the matrix
 * @param min   the minimum number
 * @param max   the maximum number
 * @param rng   the rng to use
 * @return a random matrix of the specified shape and range
 */
@Override
public INDArray rand(int[] shape, float min, float max, org.nd4j.linalg.api.rng.Random rng) {
  //ensure shapes that wind up being scalar end up with the write shape
  if (shape.length == 1 && shape[0] == 0) {
    shape = new int[] {1, 1};
  }
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray rand(int[] shape, double min, double max, org.nd4j.linalg.api.rng.Random rng) {
  Nd4j.getRandom().setSeed(rng.getSeed());
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public INDArray rand(long[] shape, double min, double max, org.nd4j.linalg.api.rng.Random rng) {
  Nd4j.getRandom().setSeed(rng.getSeed());
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: deeplearning4j/nd4j

/**
 * Create an ndarray
 * of
 * @param seed
 * @param rank
 * @param numShapes
 * @return
 */
public static int[][] getRandomBroadCastShape(long seed, int rank, int numShapes) {
  Nd4j.getRandom().setSeed(seed);
  INDArray coinFlip = Nd4j.getDistributions().createBinomial(1, 0.5).sample(new int[] {numShapes, rank});
  int[][] ret = new int[(int) coinFlip.rows()][(int) coinFlip.columns()];
  for (int i = 0; i < coinFlip.rows(); i++) {
    for (int j = 0; j < coinFlip.columns(); j++) {
      int set = coinFlip.getInt(i, j);
      if (set > 0)
        ret[i][j] = set;
      else {
        //anything from 0 to 9
        ret[i][j] = Nd4j.getRandom().nextInt(9) + 1;
      }
    }
  }
  return ret;
}

代码示例来源：origin: org.nd4j/nd4j-api

/**
 * Generates a random matrix between min and max
 *
 * @param shape the number of rows of the matrix
 * @param min   the minimum number
 * @param max   the maximum number
 * @param rng   the rng to use
 * @return a random matrix of the specified shape and range
 */
@Override
public INDArray rand(int[] shape, float min, float max, org.nd4j.linalg.api.rng.Random rng) {
  //ensure shapes that wind up being scalar end up with the write shape
  if (shape.length == 1 && shape[0] == 0) {
    shape = new int[] {1, 1};
  }
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: org.deeplearning4j/deeplearning4j-nn

@Override
public INDArray preProcess(INDArray input, int miniBatchSize) {
  return Nd4j.getDistributions().createBinomial(1, input).sample(input.shape());
}

代码示例来源：origin: org.deeplearning4j/deeplearning4j-nn

public static org.nd4j.linalg.api.rng.distribution.Distribution createDistribution(Distribution dist) {
    if (dist == null)
      return null;
    if (dist instanceof NormalDistribution) {
      NormalDistribution nd = (NormalDistribution) dist;
      return Nd4j.getDistributions().createNormal(nd.getMean(), nd.getStd());
    }
    if (dist instanceof GaussianDistribution) {
      GaussianDistribution nd = (GaussianDistribution) dist;
      return Nd4j.getDistributions().createNormal(nd.getMean(), nd.getStd());
    }
    if (dist instanceof UniformDistribution) {
      UniformDistribution ud = (UniformDistribution) dist;
      return Nd4j.getDistributions().createUniform(ud.getLower(), ud.getUpper());
    }
    if (dist instanceof BinomialDistribution) {
      BinomialDistribution bd = (BinomialDistribution) dist;
      return Nd4j.getDistributions().createBinomial(bd.getNumberOfTrials(), bd.getProbabilityOfSuccess());
    }
    throw new RuntimeException("unknown distribution type: " + dist.getClass());
  }
}

代码示例来源：origin: org.deeplearning4j/deeplearning4j-nn

/**
 * Corrupts the given input by doing a binomial sampling
 * given the corruption level
 * @param x the input to corrupt
 * @param corruptionLevel the corruption value
 * @return the binomial sampled corrupted input
 */
public INDArray getCorruptedInput(INDArray x, double corruptionLevel) {
  INDArray corrupted = Nd4j.getDistributions().createBinomial(1, 1 - corruptionLevel).sample(x.shape());
  corrupted.muli(x);
  return corrupted;
}

代码示例来源：origin: org.nd4j/nd4j-api

@Override
public INDArray rand(int[] shape, double min, double max, org.nd4j.linalg.api.rng.Random rng) {
  Nd4j.getRandom().setSeed(rng.getSeed());
  return Nd4j.getDistributions().createUniform(min, max).sample(shape);
}

代码示例来源：origin: org.nd4j/canova-nd4j-image

public static void drawMnist(DataSet mnist,INDArray reconstruct) throws InterruptedException {
  for(int j = 0; j < mnist.numExamples(); j++) {
    INDArray draw1 = mnist.get(j).getFeatureMatrix().mul(255);
    INDArray reconstructed2 = reconstruct.getRow(j);
    INDArray draw2 = Nd4j.getDistributions().createBinomial(1,reconstructed2).sample(reconstructed2.shape()).mul(255);
    DrawReconstruction d = new DrawReconstruction(draw1);
    d.title = "REAL";
    d.draw();
    DrawReconstruction d2 = new DrawReconstruction(draw2,1000,1000);
    d2.title = "TEST";
    
    d2.draw();
    Thread.sleep(1000);
    d.frame.dispose();
    d2.frame.dispose();
  }
}

代码示例来源：origin: org.datavec/datavec-data-image

public static void drawMnist(DataSet mnist, INDArray reconstruct) throws InterruptedException {
  for (int j = 0; j < mnist.numExamples(); j++) {
    INDArray draw1 = mnist.get(j).getFeatureMatrix().mul(255);
    INDArray reconstructed2 = reconstruct.getRow(j);
    INDArray draw2 = Nd4j.getDistributions().createBinomial(1, reconstructed2).sample(reconstructed2.shape())
            .mul(255);
    DrawReconstruction d = new DrawReconstruction(draw1);
    d.title = "REAL";
    d.draw();
    DrawReconstruction d2 = new DrawReconstruction(draw2, 1000, 1000);
    d2.title = "TEST";
    d2.draw();
    Thread.sleep(1000);
    d.frame.dispose();
    d2.frame.dispose();
  }
}

内容来源于网络，如有侵权，请联系作者删除！

Nd4j

相关文章

    查看更多

    热门标签

    更多
    JavaquerypythonNode开发语言requestUtil数据库Table后端算法LoggerMessageElementParser

    最新文章

    更多

    Nd4j类方法

    • 蜀ICP备13028337号-1 大数据知识库 https://www.saoniuhuo.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com