Intel® Math Kernel Library 2019 Developer Reference - C

dnnLRNCreate

Creates propagation operations for layers performing local response normalization across channels. Note: The Deep Neural Network (DNN) component in Intel MKL is deprecated and will be removed in a future release. You can continue to use optimized functions for deep neural networks through Intel Math Kernel Library for Deep Neural Networks.

Syntax

dnnError_t dnnLRNCreateForward_F32 (dnnPrimitive_t *pLrn, dnnPrimitiveAttributes_t attributes, const dnnLayout_t dataLayout, size_t kernelSie, float alpha, float beta, float k);

dnnError_t dnnLRNCreateBackward_F32 (dnnPrimitive_t *pLrn, dnnPrimitiveAttributes_t attributes, const dnnLayout_t diffLayout, const dnnLayout_t dataLayout, size_t kernelSize, float alpha, float beta, float k);

dnnError_t dnnLRNCreateForward_F64 (dnnPrimitive_t *pLrn, dnnPrimitiveAttributes_t attributes, const dnnLayout_t dataLayout, size_t kernelSize, double alpha, double beta, double k);

dnnError_t dnnLRNCreateBackward_F64 (dnnPrimitive_t *pLrn, dnnPrimitiveAttributes_t attributes, const dnnLayout_t diffLayout, const dnnLayout_t dataLayout, size_t kernelSize, double alpha, double beta, double k);

Include Files

Input Parameters

attributes

The set of attributes for the primitive.

dataLayout

The layout of the input.

diffLayout

The layout of the destination diff.

kernelSize

The number of channels to normalize across.

alpha

The α parameter.

beta

The β parameter.

k

The k parameter.

Output Parameters

pLrn

Pointer to the primitive to create:

dnnLRNCreateForward

Forward

dnnLRNCreateBackward

Backward

Description

Each dnnLRNCreate function creates a forward or backward propagation operation for local response normalization (LRN) across channels. The LRN operation is defined as: