generateWeightsHeNormal: He normal weight initialization

In darch: Package for Deep Architectures and Restricted Boltzmann Machines

Description

This function is used to generate random weights and biases using He normal weight initialization as described in He et al., http://arxiv.org/abs/1502.01852.

Usage

generateWeightsHeNormal(numUnits1, numUnits2,
  weights.mean = getParameter(".weights.mean", 0, ...), ...)

Arguments

numUnits1	Number of units in the lower layer.
numUnits2	Number of units in the upper layer.
weights.mean	mean parameter to the rnorm function.
...	Additional parameters, used for parameter resolution and passed to generateWeightsNormal.

Value

Weight matrix.

References

He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun (2015). "Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification". In: CoRR abs/1502.01852. URL: http://arxiv.org/abs/1502.01852

See Also

Other weight generation functions: generateWeightsGlorotNormal, generateWeightsGlorotUniform, generateWeightsHeUniform, generateWeightsNormal, generateWeightsUniform

Examples

## Not run: 
data(iris)
model <- darch(Species ~ ., iris, generateWeightsFunction = "generateWeightsHeNormal",
 weights.mean = .1)

## End(Not run)

darch documentation built on May 29, 2017, 8:14 p.m.