A descriptor of a pooling operation. More...
#include <dnnl_types.h>
Public Attributes | |
| dnnl_primitive_kind_t | primitive_kind |
| dnnl_prop_kind_t | prop_kind |
| dnnl_alg_kind_t | alg_kind |
| dnnl_memory_desc_t | src_desc |
| Source memory descriptor. More... | |
| dnnl_memory_desc_t | diff_src_desc |
| Source gradient memory descriptor. More... | |
| dnnl_memory_desc_t | dst_desc |
| Destination memory descriptor. More... | |
| dnnl_memory_desc_t | diff_dst_desc |
| Destination gradient memory descriptor. More... | |
| dnnl_dims_t | strides |
| Pooling kernel strides for spatial dimensions. More... | |
| dnnl_dims_t | kernel |
| Pooling kernel spatial dimensions. More... | |
| dnnl_dims_t | padding [2] |
| dnnl_data_type_t | accum_data_type |
| The accumulator data type. Initialized automatically. More... | |
Detailed Description
A descriptor of a pooling operation.
Member Data Documentation
◆ accum_data_type
| dnnl_data_type_t dnnl_pooling_desc_t::accum_data_type |
The accumulator data type. Initialized automatically.
◆ alg_kind
| dnnl_alg_kind_t dnnl_pooling_desc_t::alg_kind |
The kind of pooling algorithm. Possible values: dnnl_pooling_max, dnnl_pooling_avg_include_padding, and dnnl_pooling_avg_exclude_padding.
◆ diff_dst_desc
| dnnl_memory_desc_t dnnl_pooling_desc_t::diff_dst_desc |
Destination gradient memory descriptor.
◆ diff_src_desc
| dnnl_memory_desc_t dnnl_pooling_desc_t::diff_src_desc |
Source gradient memory descriptor.
◆ dst_desc
| dnnl_memory_desc_t dnnl_pooling_desc_t::dst_desc |
Destination memory descriptor.
◆ kernel
| dnnl_dims_t dnnl_pooling_desc_t::kernel |
Pooling kernel spatial dimensions.
◆ padding
| dnnl_dims_t dnnl_pooling_desc_t::padding[2] |
Padding in each spatial dimension. padding[0] is a padding in the beginning (padding_l), padding[1] is a padding in the end (padding_r).
◆ primitive_kind
| dnnl_primitive_kind_t dnnl_pooling_desc_t::primitive_kind |
The kind of primitive. Used for self-identifying the primitive descriptor. Must be dnnl_pooling.
◆ prop_kind
| dnnl_prop_kind_t dnnl_pooling_desc_t::prop_kind |
The kind of propagation. Possible values: dnnl_forward_training, dnnl_forward_inference, dnnl_backward, and dnnl_backward_data.
◆ src_desc
| dnnl_memory_desc_t dnnl_pooling_desc_t::src_desc |
Source memory descriptor.
◆ strides
| dnnl_dims_t dnnl_pooling_desc_t::strides |
Pooling kernel strides for spatial dimensions.
The documentation for this struct was generated from the following file:
- /work/mxnet/include/mkldnn/dnnl_types.h
