Loading TOC...

cntk:pooling-with-explicit-padding

cntk:pooling-with-explicit-padding(
   $operand as cntk:variable,
   $pooling-type as xs:string,
   $pooling-window-shape as cntk:shape,
   $strides as cntk:shape,
   $lower-pad as xs:unsignedLong*,
   $upper-pad as xs:unsignedLong*,
   [$ceil-out-dim as xs:boolean],
   [$include-pad as xs:boolean],
   [$name as xs:string]
) as cntk:function

Summary

The pooling operations compute a new tensor by selecting the maximum or average value in the pooling input. In the case of average pooling with padding, the average is only over the valid region. N-dimensional pooling allows to create max or average pooling of any dimensions, stride or padding. This function has the same functionality as cntk:pooling, with the additional ability of specify lower and upper pad values.

Parameters
$operand The operand of the operation.
$pooling-type The pooling type of the operation. Can be either "Max" or "Average".
$pooling-window-shape Dimensions of the pooling window.
$strides Strides.
$lower-pad Lower pad lengths.
$upper-pad Upper pad lengths.
$ceil-out-dim Ceiling while computing output size. Default false.
$include-pad Include pad while average pooling. Default false.
$name The name of the function instance in the network.

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