mxnet

Functions  
mkldnn_status_t MKLDNN_API  mkldnn_batch_normalization_forward_desc_init (mkldnn_batch_normalization_desc_t *bnrm_desc, mkldnn_prop_kind_t prop_kind, const mkldnn_memory_desc_t *data_desc, float epsilon, unsigned flags) 
mkldnn_status_t MKLDNN_API  mkldnn_batch_normalization_backward_desc_init (mkldnn_batch_normalization_desc_t *bnrm_desc, mkldnn_prop_kind_t prop_kind, const mkldnn_memory_desc_t *diff_data_desc, const mkldnn_memory_desc_t *data_desc, float epsilon, unsigned flags) 
A primitive to perform batch normalization.
Both forward and backward passes support inplace operation; that is, src and dst point to the same memory for forward pass, and diff_dst and diff_src point to the same memory for backward pass.
Batch normalization supports different flavors controlled by mkldnn_batch_normalization_desc_t. For example, batch normalization can compute the mean and variance on its own or take them as inputs. It can either perform scaling and shifting using gamma and beta parameters or not. Optionally, it can also perform a fused ReLU, which in case of training would also require a workspace.
mkldnn_status_t MKLDNN_API mkldnn_batch_normalization_backward_desc_init  (  mkldnn_batch_normalization_desc_t *  bnrm_desc, 
mkldnn_prop_kind_t  prop_kind,  
const mkldnn_memory_desc_t *  diff_data_desc,  
const mkldnn_memory_desc_t *  data_desc,  
float  epsilon,  
unsigned  flags  
) 
Initializes a batch normalization descriptor bnrm_desc
for backward propagation with respect to data and scaleshift parameters using memory descriptors data_desc
and diff_data_desc
, normalization parameter epsilon
, and flags
set using bit flags of type mkldnn_batch_normalization_desc_t.
Inputs:
flags
flags
Outputs:
flags
and prop_kind
= mkldnn_backwardmkldnn_status_t MKLDNN_API mkldnn_batch_normalization_forward_desc_init  (  mkldnn_batch_normalization_desc_t *  bnrm_desc, 
mkldnn_prop_kind_t  prop_kind,  
const mkldnn_memory_desc_t *  data_desc,  
float  epsilon,  
unsigned  flags  
) 
Initializes a batch normalization descriptor bnrm_desc
for forward propagation using prop_kind
(possible values are mkldnn_forward_training and mkldnn_forward_inference), memory descriptor data_desc
, normalization parameter epsilon
, and flags
set using bit flags of type mkldnn_batch_normalization_desc_t.
Inputs:
flags
flags
flags
Outputs:
flags
prop_kind
= mkldnn_forward_trainingflags
and prop_kind
= mkldnn_forward_trainingflags
and prop_kind
= mkldnn_forward_training