dnnl_eltwise_desc_t Struct Reference

A descriptor of a element-wise operation. More...

#include <dnnl_types.h>

Collaboration diagram for dnnl_eltwise_desc_t:

Public Attributes
dnnl_primitive_kind_t	primitive_kind
	The kind of primitive. More...
dnnl_prop_kind_t	prop_kind
	The kind of propagation. More...
dnnl_alg_kind_t	alg_kind
	The kind of eltwise algorithm. More...
dnnl_memory_desc_t	data_desc
	Source and destination memory descriptor.
dnnl_memory_desc_t	diff_data_desc
	Source and destination gradient memory descriptor.
float	alpha
	Algorithm specific parameter. More...

Detailed Description

A descriptor of a element-wise operation.

Examples:

cnn_inference_f32.c, cpu_cnn_training_f32.c, and cross_engine_reorder.c.

Member Data Documentation

primitive_kind

dnnl_primitive_kind_t dnnl_eltwise_desc_t::primitive_kind

The kind of primitive.

Used for self-identifying the primitive descriptor. Must be dnnl_eltwise.

prop_kind

dnnl_prop_kind_t dnnl_eltwise_desc_t::prop_kind

The kind of propagation.

Possible values: dnnl_forward_training, dnnl_forward_inference, dnnl_backward, and dnnl_backward_data.

alg_kind

dnnl_alg_kind_t dnnl_eltwise_desc_t::alg_kind

The kind of eltwise algorithm.

Possible values: dnnl_eltwise_relu, dnnl_eltwise_tanh, dnnl_eltwise_elu, dnnl_eltwise_square, dnnl_eltwise_abs, dnnl_eltwise_sqrt, dnnl_eltwise_linear, dnnl_eltwise_bounded_relu, dnnl_eltwise_soft_relu, dnnl_eltwise_logistic, dnnl_eltwise_exp, dnnl_eltwise_gelu_tanh, dnnl_eltwise_swish, dnnl_eltwise_log, dnnl_eltwise_clip, dnnl_eltwise_pow, dnnl_eltwise_gelu_erf, dnnl_eltwise_round. Possible values for passing destination memory on backward: dnnl_eltwise_relu_use_dst_for_bwd, dnnl_eltwise_tanh_use_dst_for_bwd, dnnl_eltwise_elu_use_dst_for_bwd, dnnl_eltwise_sqrt_use_dst_for_bwd, dnnl_eltwise_logistic_use_dst_for_bwd, dnnl_eltwise_exp_use_dst_for_bwd.

alpha

float dnnl_eltwise_desc_t::alpha

Algorithm specific parameter.

Accordance table:

• dnnl_eltwise_relu: alpha – negative slope, beta ignored
• dnnl_eltwise_tanh: alpha and beta ignored
• dnnl_eltwise_elu: alpha – negative slope, beta ignored
• dnnl_eltwise_square: alpha and beta ignored
• dnnl_eltwise_abs: alpha and beta ignored
• dnnl_eltwise_sqrt: alpha and beta ignored
• dnnl_eltwise_linear: alpha – scale, beta – shift
• dnnl_eltwise_bounded_relu: alpha – upper bound, beta ignored
• dnnl_eltwise_soft_relu: alpha and beta ignored
• dnnl_eltwise_logistic: alpha and beta ignored
• dnnl_eltwise_exp: alpha and beta ignored
• dnnl_eltwise_gelu_tanh: alpha and beta ignored
• dnnl_eltwise_swish: alpha – sigmoid arg scaling, beta ignored
• dnnl_eltwise_log: alpha and beta ignored
• dnnl_eltwise_clip: alpha – lower bound, beta – upper bound
• dnnl_eltwise_pow: alpha – scale, beta – exponent
• dnnl_eltwise_gelu_erf: alpha and beta ignored
• dnnl_eltwise_round: alpha and beta ignored

---

The documentation for this struct was generated from the following file:

• include/oneapi/dnnl/dnnl_types.h