Description of tensor of packed weights for rnn. More...
#include <dnnl_types.h>
Collaboration diagram for dnnl_rnn_packed_desc_t:
Public Attributes | |
| dnnl_rnn_packed_memory_format_t | format |
| int | n_parts |
| int | n |
| int | ldb |
| int | parts [DNNL_RNN_MAX_N_PARTS] |
| size_t | part_pack_size [DNNL_RNN_MAX_N_PARTS] |
| unsigned | pack_part [DNNL_RNN_MAX_N_PARTS] |
| size_t | offset_compensation |
| size_t | size |
| char | reserved [200] |
Detailed Description
Description of tensor of packed weights for rnn.
Member Data Documentation
◆ format
| dnnl_rnn_packed_memory_format_t dnnl_rnn_packed_desc_t::format |
◆ ldb
| int dnnl_rnn_packed_desc_t::ldb |
◆ n
| int dnnl_rnn_packed_desc_t::n |
◆ n_parts
| int dnnl_rnn_packed_desc_t::n_parts |
◆ offset_compensation
| size_t dnnl_rnn_packed_desc_t::offset_compensation |
◆ pack_part
| unsigned dnnl_rnn_packed_desc_t::pack_part[DNNL_RNN_MAX_N_PARTS] |
◆ part_pack_size
| size_t dnnl_rnn_packed_desc_t::part_pack_size[DNNL_RNN_MAX_N_PARTS] |
◆ parts
| int dnnl_rnn_packed_desc_t::parts[DNNL_RNN_MAX_N_PARTS] |
◆ reserved
| char dnnl_rnn_packed_desc_t::reserved[200] |
◆ size
| size_t dnnl_rnn_packed_desc_t::size |
The documentation for this struct was generated from the following file:
- /work/mxnet/include/mkldnn/oneapi/dnnl/dnnl_types.h
