Intel® oneAPI Deep Neural Network Developer Guide and Reference
A newer version of this document is available. Customers should click here to go to the newest version.
struct dnnl::group_normalization_forward::primitive_desc
Overview
Primitive descriptor for a group normalization forward propagation primitive. More…
#include <dnnl.hpp>
struct primitive_desc: public dnnl::primitive_desc
{
// construction
primitive_desc();
primitive_desc(
const engine& aengine,
prop_kind aprop_kind,
const memory::desc& src_desc,
const memory::desc& dst_desc,
memory::dim groups,
float epsilon,
normalization_flags flags,
const primitive_attr& attr = default_attr(),
bool allow_empty = false
);
primitive_desc(dnnl_primitive_desc_t pd);
// methods
memory::desc src_desc() const;
memory::desc dst_desc() const;
memory::desc weights_desc() const;
memory::desc workspace_desc() const;
memory::desc mean_desc() const;
memory::desc variance_desc() const;
dnnl::prop_kind get_prop_kind() const;
memory::dim get_group_size() const;
float get_epsilon() const;
normalization_flags get_flags() const;
};Inherited Members
public:
// methods
handle<T, traits>& operator = (const handle<T, traits>&);
handle<T, traits>& operator = (handle<T, traits>&&);
void reset(T t, bool weak = false);
T get(bool allow_empty = false) const;
operator T () const;
operator bool () const;
bool operator == (const handle<T, traits>& other) const;
bool operator != (const handle& other) const;
engine get_engine() const;
const char* impl_info_str() const;
memory::dim query_s64(query what) const;
memory::dims get_strides() const;
memory::dims get_dilations() const;
memory::dims get_padding_l() const;
memory::dims get_padding_r() const;
float get_epsilon() const;
template <typename T = unsigned>
T get_flags() const;
dnnl::algorithm get_algorithm() const;
float get_alpha() const;
float get_beta() const;
int get_axis() const;
memory::dim get_local_size() const;
float get_k() const;
float get_p() const;
std::vector<float> get_factors() const;
dnnl::algorithm get_cell_kind() const;
dnnl::rnn_direction get_direction() const;
dnnl::algorithm get_activation_kind() const;
memory::dims get_kernel() const;
memory::dim get_group_size() const;
dnnl::prop_kind get_prop_kind() const;
memory::desc query_md(query what, int idx = 0) const;
memory::desc src_desc(int idx) const;
memory::desc dst_desc(int idx) const;
memory::desc weights_desc(int idx) const;
memory::desc diff_src_desc(int idx) const;
memory::desc diff_dst_desc(int idx) const;
memory::desc diff_weights_desc(int idx) const;
memory::desc src_desc() const;
memory::desc dst_desc() const;
memory::desc weights_desc() const;
memory::desc diff_src_desc() const;
memory::desc diff_dst_desc() const;
memory::desc diff_weights_desc() const;
memory::desc workspace_desc() const;
memory::desc scratchpad_desc() const;
engine scratchpad_engine() const;
primitive_attr get_primitive_attr() const;
dnnl::primitive::kind get_kind() const;
std::vector<uint8_t> get_cache_blob_id() const;
bool next_impl();
primitive_desc_base();
primitive_desc_base();
primitive_desc_base();
primitive_desc_base();Detailed Documentation
Primitive descriptor for a group normalization forward propagation primitive.
Construction
primitive_desc()Default constructor. Produces an empty object.
primitive_desc(
const engine& aengine,
prop_kind aprop_kind,
const memory::desc& src_desc,
const memory::desc& dst_desc,
memory::dim groups,
float epsilon,
normalization_flags flags,
const primitive_attr& attr = default_attr(),
bool allow_empty = false
)Constructs a primitive descriptor for a group normalization forward propagation primitive.
Parameters:
aengine |
Engine to use. |
aprop_kind |
Propagation kind. Possible values are dnnl::prop_kind::forward_training and dnnl::prop_kind::forward_inference. |
src_desc |
Source memory descriptor. |
dst_desc |
Destination memory descriptor. |
groups |
Group normalization groups parameter. |
epsilon |
Group normalization epsilon parameter. |
flags |
Group normalization flags (dnnl::normalization_flags). |
attr |
Primitive attributes to use. Attributes are optional and default to empty attributes. |
allow_empty |
A flag signifying whether construction is allowed to fail without throwing an exception. In this case an empty object will be produced. This flag is optional and defaults to false. |
primitive_desc(dnnl_primitive_desc_t pd)Constructs a primitive descriptor for a group normalization forward propagation primitive from a C API primitive descriptor that must have a matching kind.
Parameters:
pd |
C API primitive descriptor for a group normalization forward propagation primitive. |
Methods
memory::desc src_desc() constReturns a source memory descriptor.
Returns:
Source memory descriptor.
A zero memory descriptor if the primitive does not have a source parameter.
memory::desc dst_desc() constReturns a destination memory descriptor.
Returns:
Destination memory descriptor.
A zero memory descriptor if the primitive does not have a destination parameter.
memory::desc weights_desc() constReturns a weights memory descriptor.
Returns:
Weights memory descriptor.
A zero memory descriptor if the primitive does not have a weights parameter.
memory::desc workspace_desc() constReturns the workspace memory descriptor.
Returns:
Workspace memory descriptor.
A zero memory descriptor if the primitive does not require workspace parameter.
memory::desc mean_desc() constReturns memory descriptor for mean.
Returns:
Memory descriptor for mean.
memory::desc variance_desc() constReturns memory descriptor for variance.
Returns:
Memory descriptor for variance.
dnnl::prop_kind get_prop_kind() constReturns a propagation kind.
Returns:
A propagation kind.
dnnl::prop_kind::undef if the primitive does not have a propagation parameter.
memory::dim get_group_size() constReturns a group size parameter.
Returns:
A group size parameter.
Zero if the primitive does not have a group size parameter.
float get_epsilon() constReturns an epsilon.
Returns:
An epsilon.
Zero if the primitive does not have an epsilon parameter.
normalization_flags get_flags() constReturns normalization flags.
Returns:
Normalization flags.