org.nd4j.linalg.api.ndarray.INDArray.rdivi()方法的使用及代码示例

本文整理了Java中org.nd4j.linalg.api.ndarray.INDArray.rdivi()方法的一些代码示例，展示了INDArray.rdivi()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。INDArray.rdivi()方法的具体详情如下：
包路径：org.nd4j.linalg.api.ndarray.INDArray
类名称：INDArray
方法名：rdivi

INDArray.rdivi介绍

[英]In place reverse division - i.e., (n / thisArrayValues)
[中]就地反除法-即（n/此阵列值）

代码示例

代码示例来源：origin: deeplearning4j/nd4j

/**
* Reverse division
*
* @param other the matrix to divide from
* @return
*/
@Override
public INDArray rdiv(INDArray other) {
return dup().rdivi(other);
}

代码示例来源：origin: deeplearning4j/nd4j

/**
* Reverse division
*
* @param other the matrix to subtract from
* @param result the result ndarray
* @return
*/
@Override
public INDArray rdiv(INDArray other, INDArray result) {
return dup().rdivi(other, result);
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public IComplexNDArray rdiv(IComplexNumber n, IComplexNDArray result) {
return dup().rdivi(n, result);
}

代码示例来源：origin: deeplearning4j/nd4j

@Override
public IComplexNDArray rdiv(IComplexNumber n) {
return dup().rdivi(n);
}

代码示例来源：origin: deeplearning4j/nd4j

/**
* Gets feature specific learning rates
* Adagrad keeps a history of gradients being passed in.
* Note that each gradient passed in becomes adapted over time, hence
* the opName adagrad
*
* @param gradient the gradient to get learning rates for
* @param iteration
* @return the feature specific learning rates
*/
public INDArray getGradient(INDArray gradient, int iteration) {
if (historicalGradient == null)
throw new IllegalStateException("Updater has not been initialized with view state");
historicalGradient.addi(gradient.mul(gradient));
INDArray sqrtHistory = sqrt(historicalGradient.dup(gradientReshapeOrder), false).addi(epsilon);
// lr * gradient / (sqrt(sumSquaredGradients) + epsilon)
INDArray ret = gradient.muli(sqrtHistory.rdivi(learningRate));
numIterations++;
return ret;
}

代码示例来源：origin: deeplearning4j/nd4j

/**
* Gets feature specific learning rates
* Adagrad keeps a history of gradients being passed in.
* Note that each gradient passed in becomes adapted over time, hence the opName adagrad
*
* @param gradient the gradient to get learning rates for
* @param iteration
*/
@Override
public void applyUpdater(INDArray gradient, int iteration, int epoch) {
if (historicalGradient == null)
throw new IllegalStateException("Updater has not been initialized with view state");
double learningRate = config.getLearningRate(iteration, epoch);
double epsilon = config.getEpsilon();
historicalGradient.addi(gradient.mul(gradient));
INDArray sqrtHistory = sqrt(historicalGradient.dup(gradientReshapeOrder), false).addi(epsilon);
// lr * gradient / (sqrt(sumSquaredGradients) + epsilon)
gradient.muli(sqrtHistory.rdivi(learningRate));
}
}

代码示例来源：origin: deeplearning4j/nd4j

public INDArray getGradient(INDArray gradient, int slice, int[] shape) {
boolean historicalInitialized = false;
INDArray sqrtHistory;
if (this.historicalGradient == null) {
this.historicalGradient = Nd4j.zeros(shape).add(epsilon);
historicalInitialized = true;
} else if (!this.historicalGradient.isVector()
&& this.historicalGradient.slice(slice).length() != gradient.length())
throw new IllegalArgumentException("Illegal gradient");
if (historicalGradient.isVector())
sqrtHistory = sqrt(historicalGradient);
else
sqrtHistory = !historicalInitialized ? sqrt(historicalGradient.slice(slice)) : historicalGradient;
INDArray learningRates;
try {
learningRates = sqrtHistory.rdivi(learningRate);
} catch (ArithmeticException ae) {
learningRates = sqrtHistory.rdivi(learningRate + epsilon);
}
if (gradient.length() != learningRates.length())
gradient.muli(learningRates.slice(slice));
else
gradient.muli(learningRates);
this.historicalGradient.slice(slice).addi(gradient.mul(gradient));
numIterations++;
//ensure no zeros
return gradient;
}

代码示例来源：origin: deeplearning4j/nd4j

private INDArray scoreArray(INDArray labels, INDArray preOutput, IActivation activationFn, INDArray mask) {
if (labels.size(1) != preOutput.size(1)) {
throw new IllegalArgumentException(
"Labels array numColumns (size(1) = " + labels.size(1) + ") does not match output layer"
+ " number of outputs (nOut = " + preOutput.size(1) + ") ");
}
INDArray output = activationFn.getActivation(preOutput.dup(), true);
// Clip output and labels to be between Nd4j.EPS_THREsHOLD and 1, i.e. a valid non-zero probability
output = Transforms.min(Transforms.max(output, Nd4j.EPS_THRESHOLD, false), 1, false);
labels = Transforms.min(Transforms.max(labels, Nd4j.EPS_THRESHOLD, true), 1, false);
INDArray logRatio = Transforms.log(output.rdivi(labels), false);
INDArray scoreArr = logRatio.muli(labels);
if (mask != null) {
LossUtil.applyMask(scoreArr, mask);
}
return scoreArr;
}

代码示例来源：origin: deeplearning4j/nd4j

gradient.rdivi(m).muli(alphat);

代码示例来源：origin: deeplearning4j/nd4j

INDArray dLda = output.rdivi(labels).negi();

代码示例来源：origin: deeplearning4j/nd4j

/**
* in place (element wise) division of two matrices
*
* @param other the second ndarray to divide
* @param result the result ndarray
* @return the result of the divide
*/
@Override
public INDArray divi(INDArray other, INDArray result) {
if (other.isScalar()) {
return divi(other.getDouble(0), result);
}
if (isScalar()) {
return other.rdivi(getDouble(0), result);
}
if(!Shape.shapeEquals(this.shape(),other.shape())) {
int[] broadcastDimensiOns= Shape.getBroadcastDimensions(this.shape(),other.shape());
Nd4j.getExecutioner().exec(new BroadcastDivOp(this,other,result,broadcastDimensions),broadcastDimensions);
return result;
}
LinAlgExceptions.assertSameShape(other, result);
Nd4j.getExecutioner().exec(new OldDivOp(this, other, result, length()));
if (Nd4j.ENFORCE_NUMERICAL_STABILITY)
Nd4j.clearNans(result);
return result;
}

代码示例来源：origin: improbable-research/keanu

@Override
public DoubleTensor reciprocalInPlace() {
tensor.rdivi(1.0);
return this;
}

代码示例来源：origin: org.nd4j/nd4j-api

/**
* Reverse division
*
* @param other the matrix to divide from
* @return
*/
@Override
public INDArray rdiv(INDArray other) {
return dup().rdivi(other);
}

代码示例来源：origin: org.nd4j/nd4j-api

/**
* Reverse division
*
* @param other the matrix to subtract from
* @param result the result ndarray
* @return
*/
@Override
public INDArray rdiv(INDArray other, INDArray result) {
return dup().rdivi(other, result);
}

代码示例来源：origin: org.nd4j/nd4j-api

@Override
public IComplexNDArray rdiv(IComplexNumber n) {
return dup().rdivi(n);
}

代码示例来源：origin: org.nd4j/nd4j-api

@Override
public IComplexNDArray rdiv(IComplexNumber n, IComplexNDArray result) {
return dup().rdivi(n, result);
}

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

@Override
public INDArray generateAtMean(INDArray preOutDistributionParams) {
//Input: gamma = log(lambda) -> lambda = exp(gamma)
//Mean for exponential distribution: 1/lambda
INDArray gamma = activationFn.getActivation(preOutDistributionParams.dup(), false);
INDArray lambda = Transforms.exp(gamma, true);
return lambda.rdivi(1.0); //mean = 1.0 / lambda
}

代码示例来源：origin: org.deeplearning4j/deeplearning4j-cuda-9.2

@Override
public INDArray getVarCache() {
INDArray ret;
if(Nd4j.dataType() == DataBuffer.Type.HALF){
INDArray vc = varCache.convertToHalfs();
ret = vc.mul(vc).rdivi(1.0).subi(eps);
} else {
ret = varCache.mul(varCache).rdivi(1.0).subi(eps);
}
if(Nd4j.dataType() == DataBuffer.Type.HALF){
//Buffer is FP32
return ret.convertToHalfs();
}
return ret;
}

代码示例来源：origin: org.nd4j/nd4j-api

/**
* Gets feature specific learning rates
* Adagrad keeps a history of gradients being passed in.
* Note that each gradient passed in becomes adapted over time, hence the name adagrad
*
* @param gradient the gradient to get learning rates for
* @param iteration
*/
@Override
public void applyUpdater(INDArray gradient, int iteration) {
if (historicalGradient == null)
throw new IllegalStateException("Updater has not been initialized with view state");
double learningRate = config.getLearningRate();
double epsilon = config.getEpsilon();
historicalGradient.addi(gradient.mul(gradient));
INDArray sqrtHistory = sqrt(historicalGradient.dup(gradientReshapeOrder), false).addi(epsilon);
// lr * gradient / (sqrt(sumSquaredGradients) + epsilon)
gradient.muli(sqrtHistory.rdivi(learningRate));
}
}

代码示例来源：origin: org.nd4j/nd4j-api

private INDArray scoreArray(INDArray labels, INDArray preOutput, IActivation activationFn, INDArray mask) {
if (labels.size(1) != preOutput.size(1)) {
throw new IllegalArgumentException("Labels array numColumns (size(1) = " + labels.size(1)
+ ") does not match output layer" + " number of outputs (nOut = " + preOutput.size(1)
+ ") ");

}
INDArray output = activationFn.getActivation(preOutput.dup(), true);
// Clip output and labels to be between Nd4j.EPS_THREsHOLD and 1, i.e. a valid non-zero probability
output = Transforms.min(Transforms.max(output, Nd4j.EPS_THRESHOLD, false), 1, false);
labels = Transforms.min(Transforms.max(labels, Nd4j.EPS_THRESHOLD, true), 1, false);
INDArray logRatio = Transforms.log(output.rdivi(labels), false);
INDArray scoreArr = logRatio.muli(labels);
if (mask != null) {
LossUtil.applyMask(scoreArr, mask);
}
return scoreArr;
}