Developer Guide for Intel® Data Analytics Acceleration Library 2018
The backward two-dimensional max pooling layer accepts the input described below. Pass the Input ID as a parameter to the methods that provide input for your algorithm. For more details, see Algorithms.
Input ID |
Input |
|
---|---|---|
inputGradient |
Pointer to tensor G of size l1 x ... x lp that stores the input gradient computed on the preceding layer. This input can be an object of any class derived from Tensor. |
|
inputFromForward |
Collection of data needed for the backward two-dimensional max pooling layer. |
|
Element ID |
Element |
|
auxSelectedIndices |
Tensor T of size l1 x ... x lp that stores indices of maximum elements. |
|
auxInputDimensions |
NumericTable of size 1 x p that stores the sizes of the dimensions of input data tensor X: n1, n2, …, np. |
For common parameters of neural network layers, see Common Parameters.
In addition to the common parameters, the backward two-dimensional max pooling layer has the following parameters:
Parameter |
Default Value |
Description |
|
---|---|---|---|
algorithmFPType |
float |
The floating-point type that the algorithm uses for intermediate computations. Can be float or double. |
|
method |
defaultDense |
Performance-oriented computation method, the only method supported by the layer. |
|
kernelSizes |
KernelSizes(2, 2) |
Data structure representing the size of the two-dimensional subtensor from which the maximum element is selected. |
|
strides |
Strides(2, 2) |
Data structure representing intervals s1, s2 on which the subtensors for max pooling are selected. |
|
paddings |
Paddings(0, 0) |
Data structure representing the number of data elements to implicitly add to each side of the two-dimensional subtensor along which max pooling is performed. |
|
indices |
Indices(p-2, p-1) |
Indices k1, k2 of the dimensions along which max pooling is performed. |
The backward two-dimensional max pooling layer calculates the result described below. Pass the Result ID as a parameter to the methods that access the results of your algorithm. For more details, see Algorithms.
Result ID |
Result |
|
---|---|---|
gradient |
Pointer to tensor Z of size n1 x ... x np that stores the result of the backward two-dimensional max pooling layer. This input can be an object of any class derived from Tensor. |
C++: max_pool2d_layer_dense_batch.cpp
Java*: MaxPool2DLayerDenseBatch.java
Python*: max_pool2d_layer_dense_batch.py