A descriptor of a Layer Normalization operation. More...
#include <dnnl_types.h>
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_memory_desc_t | data_desc |
Source and destination memory descriptor. | |
dnnl_memory_desc_t | diff_data_desc |
Source and destination gradient memory descriptor. | |
dnnl_memory_desc_t | data_scaleshift_desc |
Scale and shift data and gradient memory descriptors. More... | |
dnnl_memory_desc_t | stat_desc |
Mean and variance data memory descriptors. More... | |
float | layer_norm_epsilon |
Layer normalization epsilon parameter. | |
A descriptor of a Layer Normalization operation.
dnnl_primitive_kind_t dnnl_layer_normalization_desc_t::primitive_kind |
The kind of primitive.
Used for self-identifying the primitive descriptor. Must be dnnl_layer_normalization.
dnnl_prop_kind_t dnnl_layer_normalization_desc_t::prop_kind |
The kind of propagation.
Possible values: dnnl_forward_training, dnnl_forward_inference, dnnl_backward, and dnnl_backward_data.
dnnl_memory_desc_t dnnl_layer_normalization_desc_t::data_scaleshift_desc |
Scale and shift data and gradient memory descriptors.
Scaleshift memory descriptor uses 2D dnnl_ab format[2, normalized_dim] where 1-st dimension contains gamma parameter, 2-nd dimension contains beta parameter. Normalized_dim is equal to the last logical dimension of the data tensor across which normalization is performed.
dnnl_memory_desc_t dnnl_layer_normalization_desc_t::stat_desc |
Mean and variance data memory descriptors.
Statistics (mean and variance) memory descriptor is the k-dimensional tensor where k is equal to data_tensor_ndims - 1 and may have any plain (stride[last_dim] == 1) user-provided format.