mxnet
|
#include <mkldnn_types.h>
Public Attributes | |
int | ndims |
Number of dimensions. More... | |
mkldnn_dims_t | dims |
mkldnn_data_type_t | data_type |
Data type of the tensor elements. More... | |
mkldnn_dims_t | padded_dims |
Size of the data including padding in each dimension. More... | |
mkldnn_dims_t | padded_offsets |
mkldnn_dim_t | offset0 |
mkldnn_format_kind_t | format_kind |
Memory format kind. More... | |
union { | |
mkldnn_blocking_desc_t blocking | |
mkldnn_wino_desc_t wino_desc | |
Tensor of weights for integer 8bit winograd convolution. More... | |
mkldnn_rnn_packed_desc_t rnn_packed_desc | |
Tensor of packed weights for RNN. More... | |
} | format_desc |
mkldnn_memory_extra_desc_t | extra |
Memory descriptor. The description is based on a number of dimensions, dimensions themselves, plus information about elements type and memory format. Additionally, contains format-specific descriptions of the data layout.
mkldnn_blocking_desc_t mkldnn_memory_desc_t::blocking |
Description of the data layout for memory formats that use blocking.
mkldnn_data_type_t mkldnn_memory_desc_t::data_type |
Data type of the tensor elements.
mkldnn_dims_t mkldnn_memory_desc_t::dims |
Dimensions in the following order:
{N, C, [[D,] H,] W}
){[G,] O, I, [[D,] H,] W}
){T, N, C}
) or layers, directions, states, mini-batch, channels ({L, D, S, N, C}
){L, D, I, G, O}
).{N, C, H, W}
. mkldnn_memory_extra_desc_t mkldnn_memory_desc_t::extra |
union { ... } mkldnn_memory_desc_t::format_desc |
mkldnn_format_kind_t mkldnn_memory_desc_t::format_kind |
Memory format kind.
int mkldnn_memory_desc_t::ndims |
Number of dimensions.
mkldnn_dim_t mkldnn_memory_desc_t::offset0 |
Offset from memory origin to the current block, non-zero only in a description of a memory sub-block.
mkldnn_dims_t mkldnn_memory_desc_t::padded_dims |
Size of the data including padding in each dimension.
mkldnn_dims_t mkldnn_memory_desc_t::padded_offsets |
Per-dimension offset from the padding to actual data, the top-level tensor with offsets applied must lie within the padding area.
mkldnn_rnn_packed_desc_t mkldnn_memory_desc_t::rnn_packed_desc |
Tensor of packed weights for RNN.
mkldnn_wino_desc_t mkldnn_memory_desc_t::wino_desc |
Tensor of weights for integer 8bit winograd convolution.