struct dnnl_batch_normalization_desc_t¶
Overview¶
A descriptor of a Batch Normalization operation. More…
#include <dnnl_types.h> struct dnnl_batch_normalization_desc_t { // fields dnnl_primitive_kind_t primitive_kind; dnnl_prop_kind_t prop_kind; dnnl_memory_desc_t data_desc; dnnl_memory_desc_t diff_data_desc; dnnl_memory_desc_t data_scaleshift_desc; dnnl_memory_desc_t diff_data_scaleshift_desc; dnnl_memory_desc_t stat_desc; float batch_norm_epsilon; unsigned flags; };
Detailed Documentation¶
A descriptor of a Batch Normalization operation.
Fields¶
dnnl_primitive_kind_t primitive_kind
The kind of primitive.
Used for self-identifying the primitive descriptor. Must be dnnl_batch_normalization.
dnnl_prop_kind_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 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.
Scaleshift memory descriptor uses 2D dnnl_nc format[2,Channels]. 1-st dimension contains gamma parameter, 2-nd dimension contains beta parameter.
dnnl_memory_desc_t stat_desc
Statistics memory descriptor.
Statistics (mean or variance) descriptor use 1D dnnl_x format[Channels].
float batch_norm_epsilon
Batch normalization epsilon parameter.