Loading TOC...

cntk:mean-variance-normalization

cntk:mean-variance-normalization(
   $operand as cntk:variable,
   $epsilon as xs:double,
   [$use-stats-across-channels as xs:boolean],
   [$do-variance-scaling as xs:boolean],
   [$name as xs:string]
) as cntk:function

Summary

Computes mean-variance normalization of the specified input operand. This operation computes and mean and variance for the entire tensor if use_stats_across_channels is True. If use_stats_across_channels is False the computes mean and variance per channel and normalizes each channel with its own mean and variance. If do_variance_scaling is False, only the mean is subtracted, and the variance scaling is omitted.

Parameters
$operand Input tensor.
$epsilon Epsilon added to the standard deviation to avoid division by 0.
$use-stats-across-channels If False, mean and variance are computed per channel. If True, mean and variance are computed over the entire tensor (all axes).
$do-variance-scaling If False, only the mean is subtracted. If True, it is also scaled by inverse of standard deviation.
$name The name of the function instance in the network.

Example

  let $input-variable1 := cntk:input-variable(cntk:shape((3)), "float",
    fn:false(), fn:false(), "feature")
  let $input-variable2 := cntk:input-variable(cntk:shape((3)), "float",
    fn:false(), fn:false(), "feature")
  return cntk:mean-variance-normalization($input-variable2, 281775769.568754,
    fn:true(), fn:false(), "4RM")
  => cntk:function(Composite MeanVarianceNormalization (Input(Name(feature),
  Shape([3]), Dynamic Axes([Sequence Axis(Default Dynamic Axis), Batch Axis(
  Default Batch Axis)]))))

Stack Overflow iconStack Overflow: Get the most useful answers to questions from the MarkLogic community, or ask your own question.