Pooling

API Reference

The pooling primitive performs forward or backward max or average pooling operation on 1D, 2D, or 3D spatial data.

The pooling operation is defined by the following formulas. We show formulas only for 2D spatial data which are straightforward to generalize to cases of higher and lower dimensions. Variable names follow the standard Naming Conventions.

Forward

Max pooling:

\[ \dst(n, c, oh, ow) = \max\limits_{kh, kw} \left( \src(n, c, oh \cdot SH + kh - PH_L, ow \cdot SW +kw - PW_L) \right) \]

Average pooling:

\[ \dst(n, c, oh, ow) = \frac{1}{DENOM} \sum\limits_{kh, kw} \src(n, c, oh \cdot SH + kh - PH_L, ow \cdot SW +kw - PW_L) \]

Here output spatial dimensions are calculated similarly to how they are done in Convolution.

Average pooling supports two algorithms:

• dnnl_pooling_avg_include_padding, in which case \(DENOM = KH \cdot KW\),
• dnnl_pooling_avg_exclude_padding, in which case \(DENOM\) equals to the size of overlap between an averaging window and images.

TODO: a picture would be nice here.

Difference Between Forward Training and Forward Inference

• Max pooling requires a workspace for the dnnl_forward_training propagation kind, and doesn't require it for dnnl_forward_inference (see details below).

Backward

The backward propagation computes \(\)\diffsrc\f$(n, c, h, w) \(, based on \) \(\diffdst\)(n, c, h, w) \( and (in case of max pooling) <tt>workspace</tt>. @section autotoc_md232 Execution Arguments When executed, the inputs and outputs should be mapped to an execution argument index as specified by the following table. <table class="markdownTable"> <tr class="markdownTableHead"> <th class="markdownTableHeadNone"> Primitive input/output \ilinebr </th> <th class="markdownTableHeadNone"> Execution argument index \ilinebr </th> </tr> <tr class="markdownTableRowOdd"> <td class="markdownTableBodyNone"> \)\src\f$

DNNL_ARG_SRC

\(\dst\)

DNNL_ARG_DST

workspace

DNNL_ARG_WORKSPACE

\(\diffsrc\)

DNNL_ARG_DIFF_SRC

\(\diffdst\)

DNNL_ARG_DIFF_DST

Implementation Details

General Notes

1. During training, max pooling requires a workspace on forward (dnnl_forward_training) and backward passes to save indices where a maximum was found. The workspace format is opaque, and the indices cannot be restored from it. However, one can use backward pooling to perform up-sampling (used in some detection topologies). The workspace can be created via workspace_desc() from the pooling primitive descriptor.
2. A user can use memory format tag dnnl_format_tag_any for dst memory descriptor when creating pooling forward propagation. The library would derive the appropriate format from the src memory descriptor. However, the src itself must be defined. Similarly, a user can use memory format tag dnnl_format_tag_any for the diff_src memory descriptor when creating pooling backward propagation.

Data Type Support

The pooling primitive supports the following combinations of data types:

Propagation	Source / Destination	Accumulation data type (used for average pooling only)
forward / backward	f32, bf16	f32
forward	f16	f16
forward	s8, u8, s32	s32

Warning

There might be hardware and/or implementation specific restrictions. Check Implementation Limitations section below.

Data Representation

Source, Destination, and Their Gradients

Like other CNN primitives, the pooling primitive expects data to be an \(N \times C \times W\) tensor for the 1D spatial case, an \(N \times C \times H \times W\) tensor for the 2D spatial case, and an \(N \times C \times D \times H \times W\) tensor for the 3D spatial case.

The pooling primitive is optimized for the following memory formats:

Spatial	Logical tensor	Data type	Implementations optimized for memory formats
1D	NCW	f32	dnnl_ncw (dnnl_abc), dnnl_nwc (dnnl_acb), optimized^
1D	NCW	s32, s8, u8	dnnl_nwc (dnnl_acb), optimized^
2D	NCHW	f32	dnnl_nchw (dnnl_abcd), dnnl_nhwc (dnnl_acdb), optimized^
2D	NCHW	s32, s8, u8	dnnl_nhwc (dnnl_acdb), optimized^
3D	NCDHW	f32	dnnl_ncdhw (dnnl_abcde), dnnl_ndhwc (dnnl_acdeb), optimized^
3D	NCDHW	s32, s8, u8	dnnl_ndhwc (dnnl_acdeb), optimized^

Here optimized^ means the format that comes out of any preceding compute-intensive primitive.

Post-ops and Attributes

The pooling primitive does not support any post-ops or attributes.

Implementation Limitations

1. No primitive specific limitations. Refer to Data Types for limitations related to data types support.

Performance Tips

N/A

Examples

Engine	Name	Comments
CPU/GPU	Pooling Primitive Example	This C++ API example demonstrates how to create and execute a Pooling primitive in forward training propagation mode.