LRN

Overview

A primitive to perform local response normalization (LRN) across or within channels. More…

// structs

struct dnnl_lrn_desc_t;
struct dnnl::lrn_backward;
struct dnnl::lrn_forward;

// global functions

dnnl_status_t DNNL_API dnnl_lrn_forward_desc_init(
    dnnl_lrn_desc_t* lrn_desc,
    dnnl_prop_kind_t prop_kind,
    dnnl_alg_kind_t alg_kind,
    const dnnl_memory_desc_t* data_desc,
    dnnl_dim_t local_size,
    float alpha,
    float beta,
    float k
    );

dnnl_status_t DNNL_API dnnl_lrn_backward_desc_init(
    dnnl_lrn_desc_t* lrn_desc,
    dnnl_alg_kind_t alg_kind,
    const dnnl_memory_desc_t* diff_data_desc,
    const dnnl_memory_desc_t* data_desc,
    dnnl_dim_t local_size,
    float alpha,
    float beta,
    float k
    );

Detailed Documentation

A primitive to perform local response normalization (LRN) across or within channels.

See also:

Local Response Normalization (LRN) in developer guide

Global Functions

dnnl_status_t DNNL_API dnnl_lrn_forward_desc_init(
    dnnl_lrn_desc_t* lrn_desc,
    dnnl_prop_kind_t prop_kind,
    dnnl_alg_kind_t alg_kind,
    const dnnl_memory_desc_t* data_desc,
    dnnl_dim_t local_size,
    float alpha,
    float beta,
    float k
    )

Initializes a descriptor for LRN forward propagation primitive.

Parameters:

lrn_desc

Output descriptor for a LRN primitive.

prop_kind

Propagation kind. Possible values are dnnl_forward_training and dnnl_forward_inference.

alg_kind

LRN algorithm kind: either dnnl_lrn_across_channels or dnnl_lrn_within_channel.

data_desc

Source and destination memory descriptor.

local_size

Regularization local size.

alpha

The alpha regularization parameter.

beta

The beta regularization parameter.

k

The k regularization parameter.

Returns:

dnnl_success on success and a status describing the error otherwise.

dnnl_status_t DNNL_API dnnl_lrn_backward_desc_init(
    dnnl_lrn_desc_t* lrn_desc,
    dnnl_alg_kind_t alg_kind,
    const dnnl_memory_desc_t* diff_data_desc,
    const dnnl_memory_desc_t* data_desc,
    dnnl_dim_t local_size,
    float alpha,
    float beta,
    float k
    )

Initializes a descriptor for LRN backward propagation primitive.

Parameters:

lrn_desc

Output descriptor for a LRN primitive.

alg_kind

LRN algorithm kind: either dnnl_lrn_across_channels or dnnl_lrn_within_channel.

diff_data_desc

Diff source and diff destination memory descriptor.

data_desc

Source memory descriptor.

local_size

Regularization local size.

alpha

The alpha regularization parameter.

beta

The beta regularization parameter.

k

The k regularization parameter.

Returns:

dnnl_success on success and a status describing the error otherwise.