Developer Guide for Intel® Data Analytics Acceleration Library 2018 Update 2
The backward concat 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 the tensor of size n1 x ... x np that stores input gradient G 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 concat layer. |
|
Element ID |
Element |
|
auxInputDimensions |
Collection of integers that stores the sizes along the k-th dimension of the input tensors in the forward computation step: nk1, ..., nkt. |
For common parameters of neural network layers, see Common Parameters.
In addition to the common parameters, the backward concat layer has the following parameters:
Parameter |
Default Value |
Description |
|
---|---|---|---|
concatDimension |
0 |
Index of the dimension along which deconcatenation should be implemented. |
The backward concat 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 |
|
---|---|---|
resultLayerData |
Collection of tensors of size n1 x ... x nkj x ... x np that stores the result of the backward concat layer. This collection can contain objects of any class derived from Tensor. |
The gradient field stores a null pointer. All the computation results are stored in resultLayerData.
C++: concat_layer_dense_batch.cpp
Java*: ConcatLayerDenseBatch.java
Python*: concat_layer_dense_batch.py