Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN)
0.10
Performance library for Deep Learning
|
A descriptor of a element-wise operation. More...
#include <mkldnn_types.h>
Public Attributes | |
mkldnn_primitive_kind_t | primitive_kind |
The kind of primitive. More... | |
mkldnn_prop_kind_t | prop_kind |
The kind of propagation. More... | |
mkldnn_alg_kind_t | alg_kind |
The kind of eltwise algorithm. More... | |
mkldnn_memory_desc_t | data_desc |
Source and destination memory descriptor. More... | |
mkldnn_memory_desc_t | diff_data_desc |
Source and destination gradient memory descriptor. More... | |
double | alpha |
Algorithm specific parameter. More... | |
double | beta |
double | negative_slope |
Scaling factor for negative values. More... | |
A descriptor of a element-wise operation.
mkldnn_alg_kind_t mkldnn_eltwise_desc_t::alg_kind |
The kind of eltwise algorithm.
Possible values: mkldnn_eltwise_relu, mkldnn_eltwise_tanh, mkldnn_eltwise_elu
double mkldnn_eltwise_desc_t::alpha |
Algorithm specific parameter.
Accordance table:
alpha
– negative slope, beta
ignoredalpha
and beta
ignoredalpha
– negative slope, beta
ignored double mkldnn_eltwise_desc_t::beta |
mkldnn_memory_desc_t mkldnn_eltwise_desc_t::data_desc |
Source and destination memory descriptor.
mkldnn_memory_desc_t mkldnn_eltwise_desc_t::diff_data_desc |
Source and destination gradient memory descriptor.
double mkldnn_eltwise_desc_t::negative_slope |
Scaling factor for negative values.
Stored as double-precision, but interpreted in a way specific to the data type in each implementation.
mkldnn_primitive_kind_t mkldnn_eltwise_desc_t::primitive_kind |
The kind of primitive.
Used for self identifying the primitive descriptor. Must be mkldnn_relu.
mkldnn_prop_kind_t mkldnn_eltwise_desc_t::prop_kind |
The kind of propagation.
Possible values: mkldnn_forward_training, mkldnn_forward_inference, mkldnn_backward, and mkldnn_backward_data.