Deep Neural Network Library (DNNL)  1.1.3
Performance library for Deep Learning
Functions

A primitive to perform softmax. More...

Functions

dnnl_status_t DNNL_API dnnl_softmax_forward_desc_init (dnnl_softmax_desc_t *softmax_desc, dnnl_prop_kind_t prop_kind, const dnnl_memory_desc_t *data_desc, int softmax_axis)
 Initializes a softmax_desc for forward propagation using prop_kind (possible values are dnnl_forward_training and dnnl_forward_inference) and memory descriptor data_desc. More...
 
dnnl_status_t DNNL_API dnnl_softmax_backward_desc_init (dnnl_softmax_desc_t *softmax_desc, const dnnl_memory_desc_t *diff_desc, const dnnl_memory_desc_t *data_desc, int softmax_axis)
 Initializes a softmax_desc for backward propagation using memory descriptors diff_desc and data_desc. More...
 

Detailed Description

A primitive to perform softmax.

See also
Softmax in developer guide
Softmax in C++ API

Function Documentation

◆ dnnl_softmax_forward_desc_init()

dnnl_status_t DNNL_API dnnl_softmax_forward_desc_init ( dnnl_softmax_desc_t softmax_desc,
dnnl_prop_kind_t  prop_kind,
const dnnl_memory_desc_t data_desc,
int  softmax_axis 
)

Initializes a softmax_desc for forward propagation using prop_kind (possible values are dnnl_forward_training and dnnl_forward_inference) and memory descriptor data_desc.

Inputs:

Outputs:

◆ dnnl_softmax_backward_desc_init()

dnnl_status_t DNNL_API dnnl_softmax_backward_desc_init ( dnnl_softmax_desc_t softmax_desc,
const dnnl_memory_desc_t diff_desc,
const dnnl_memory_desc_t data_desc,
int  softmax_axis 
)

Initializes a softmax_desc for backward propagation using memory descriptors diff_desc and data_desc.

Inputs:

Outputs: