.rst .pdf

Contents

Resampling

Overview

A primitive to compute resampling operation on 1D, 2D or 3D data tensor using Nearest Neighbor, or Linear (Bilinear, Trilinear) interpolation method. More…

// structs

struct dnnl_resampling_desc_t;
struct dnnl::resampling_backward;
struct dnnl::resampling_forward;

// global functions

dnnl_status_t DNNL_API dnnl_resampling_forward_desc_init(
    dnnl_resampling_desc_t* resampling_desc,
    dnnl_prop_kind_t prop_kind,
    dnnl_alg_kind_t alg_kind,
    const float* factors,
    const dnnl_memory_desc_t* src_desc,
    const dnnl_memory_desc_t* dst_desc
    );

dnnl_status_t DNNL_API dnnl_resampling_backward_desc_init(
    dnnl_resampling_desc_t* resampling_desc,
    dnnl_alg_kind_t alg_kind,
    const float* factors,
    const dnnl_memory_desc_t* diff_src_desc,
    const dnnl_memory_desc_t* diff_dst_desc
    );

Detailed Documentation

A primitive to compute resampling operation on 1D, 2D or 3D data tensor using Nearest Neighbor, or Linear (Bilinear, Trilinear) interpolation method.

See also:

Resampling in developer guide

Global Functions

dnnl_status_t DNNL_API dnnl_resampling_forward_desc_init(
    dnnl_resampling_desc_t* resampling_desc,
    dnnl_prop_kind_t prop_kind,
    dnnl_alg_kind_t alg_kind,
    const float* factors,
    const dnnl_memory_desc_t* src_desc,
    const dnnl_memory_desc_t* dst_desc
    )

Initializes a descriptor for a resampling forward propagation primitive.

Note

Destination memory descriptor is allowed to be initialized with dnnl_format_tag_any or with format_kind set to dnnl_format_kind_any.

Parameters:

resampling_desc	Output descriptor for a resampling primitive.
prop_kind	Propagation kind. Possible values are dnnl_forward_training and dnnl_forward_inference.
alg_kind	resampling algorithm kind: either dnnl_resampling_nearest, or dnnl_resampling_linear.
factors	Array of scaling factors for spatial dimension.
src_desc	Source memory descriptor.
dst_desc	Destination memory descriptor.

Returns:

dnnl_success on success and a status describing the error otherwise.

dnnl_status_t DNNL_API dnnl_resampling_backward_desc_init(
    dnnl_resampling_desc_t* resampling_desc,
    dnnl_alg_kind_t alg_kind,
    const float* factors,
    const dnnl_memory_desc_t* diff_src_desc,
    const dnnl_memory_desc_t* diff_dst_desc
    )

Initializes a descriptor for resampling backward propagation primitive.

Parameters:

resampling_desc	Output descriptor for a resampling primitive.
alg_kind	resamplinging algorithm kind: either dnnl_resampling_nearest, or dnnl_resampling_linear.
diff_src_desc	Diff source memory descriptor.
diff_dst_desc	Diff destination memory descriptor.
factors	Array of scaling factors for spatial dimension.

Returns:

dnnl_success on success and a status describing the error otherwise.

struct dnnl::matmul::primitive_desc struct dnnl_resampling_desc_t