| Package | Description |
|---|---|
| org.bytedeco.javacpp |
| Modifier and Type | Method and Description |
|---|---|
mkldnn.batch_normalization_forward.desc |
mkldnn.batch_normalization_forward.desc.data(mkldnn.mkldnn_batch_normalization_desc_t data) |
mkldnn.batch_normalization_backward.desc |
mkldnn.batch_normalization_backward.desc.data(mkldnn.mkldnn_batch_normalization_desc_t data) |
static int |
mkldnn.mkldnn_batch_normalization_backward_desc_init(mkldnn.mkldnn_batch_normalization_desc_t bnrm_desc,
int prop_kind,
mkldnn.mkldnn_memory_desc_t diff_data_desc,
mkldnn.mkldnn_memory_desc_t data_desc,
float epsilon,
int flags)
Initializes a batch normalization descriptor \p bnrm_desc for backward
propagation with respect to data and scale-shift parameters using memory
descriptors \p data_desc and \p diff_data_desc, and normalization parameter
\p epsilon and flags (possible values are #mkldnn_use_global_stats and
#mkldnn_use_scaleshift).
|
static int |
mkldnn.mkldnn_batch_normalization_forward_desc_init(mkldnn.mkldnn_batch_normalization_desc_t bnrm_desc,
int prop_kind,
mkldnn.mkldnn_memory_desc_t data_desc,
float epsilon,
int flags)
Initializes a batch normalization descriptor \p bnrm_desc for forward
propagation using \p prop_kind, (possible values are
#mkldnn_forward_training or #mkldnn_forward_inference), memory descriptor
\p data_desc, normalization parameter \p epsilon and flags (possible values
are #mkldnn_use_global_stats and #mkldnn_use_scaleshift).
|
Copyright © 2018. All rights reserved.