.rst .pdf

Contents

Batch Normalization

Overview

A primitive to perform batch normalization. More…

// structs

struct dnnl::batch_normalization_backward;
struct dnnl::batch_normalization_forward;
struct dnnl_batch_normalization_desc_t;

// global functions

dnnl_status_t DNNL_API dnnl_batch_normalization_forward_desc_init(
    dnnl_batch_normalization_desc_t* bnrm_desc,
    dnnl_prop_kind_t prop_kind,
    const dnnl_memory_desc_t* data_desc,
    float epsilon,
    unsigned flags
    );

dnnl_status_t DNNL_API dnnl_batch_normalization_backward_desc_init(
    dnnl_batch_normalization_desc_t* bnrm_desc,
    dnnl_prop_kind_t prop_kind,
    const dnnl_memory_desc_t* diff_data_desc,
    const dnnl_memory_desc_t* data_desc,
    float epsilon,
    unsigned flags
    );

Detailed Documentation

A primitive to perform batch normalization.

Both forward and backward propagation primitives support in-place operation; that is, src and dst can refer to the same memory for forward propagation, and diff_dst and diff_src can refer to the same memory for backward propagation.

The batch normalization primitives computations can be controlled by specifying different dnnl::normalization_flags values. For example, batch normalization can compute the mean and variance on its own or take them as inputs. It can either perform scaling and shifting using gamma and beta parameters or not. Optionally, it can also perform a fused ReLU, which in case of training would also require a workspace.

See also:

Batch Normalization in developer guide

Global Functions

dnnl_status_t DNNL_API dnnl_batch_normalization_forward_desc_init(
    dnnl_batch_normalization_desc_t* bnrm_desc,
    dnnl_prop_kind_t prop_kind,
    const dnnl_memory_desc_t* data_desc,
    float epsilon,
    unsigned flags
    )

Initializes a descriptor for a batch normalization forward propagation primitive.

Note

In-place operation is supported: the dst can refer to the same memory as the src.

Parameters:

bnrm_desc	Output descriptor for batch normalization primitive.
prop_kind	Propagation kind. Possible values are dnnl_forward_training and dnnl_forward_inference.
data_desc	Source and destination memory descriptor.
epsilon	Batch normalization epsilon parameter.
flags	Batch normalization flags (dnnl_normalization_flags_t).

Returns:

dnnl_success on success and a status describing the error otherwise.

dnnl_status_t DNNL_API dnnl_batch_normalization_backward_desc_init(
    dnnl_batch_normalization_desc_t* bnrm_desc,
    dnnl_prop_kind_t prop_kind,
    const dnnl_memory_desc_t* diff_data_desc,
    const dnnl_memory_desc_t* data_desc,
    float epsilon,
    unsigned flags
    )

Initializes a descriptor for a batch normalization backward propagation primitive.

Note

In-place operation is supported: the diff_dst can refer to the same memory as the diff_src.

Parameters:

bnrm_desc	Output descriptor for batch normalization primitive.
prop_kind	Propagation kind. Possible values are dnnl_backward_data and dnnl_backward (diffs for all parameters are computed in this case).
diff_data_desc	Diff source and diff destination memory descriptor.
data_desc	Source memory descriptor.
epsilon	Batch normalization epsilon parameter.
flags	Batch normalization flags (dnnl_normalization_flags_t).

Returns:

dnnl_success on success and a status describing the error otherwise.

struct dnnl::lrn_forward::primitive_desc struct dnnl::batch_normalization_backward