cldnn::fully_connected Struct Reference

Performs forward fully connected layer (inner product). Also supports built-in Relu cldnn_activation_desc available by setting it in arguments. . More...

#include <fully_connected.hpp>

Inheritance diagram for cldnn::fully_connected:

Collaboration diagram for cldnn::fully_connected:

Public Types
typedef cldnn_fully_connected_desc	dto

Public Member Functions
	fully_connected (const primitive_id &id, const primitive_id &input, const primitive_id &weights, const primitive_id &bias="", bool with_activation=false, float activation_slp=0.0f, const padding &output_padding=padding())
	Constructs fully connected layer. More...
	fully_connected (const dto *dto)
	Constructs a copy from basic C API cldnn_fully_connected_desc.
Public Member Functions inherited from cldnn::primitive_base< fully_connected, cldnn_fully_connected_desc >
const cldnn_primitive_desc *	get_dto () const override
	Returns pointer to a C API primitive descriptor casted to cldnn_primitive_desc.
Public Member Functions inherited from cldnn::primitive
	primitive (const primitive_type_id &type, const primitive_id &id, const std::vector< primitive_id > &input, const padding &output_padding=padding())
	primitive (const cldnn_primitive_desc *dto)
	Constructs a copy from basic C API cldnn_primitive_desc.
std::vector< std::reference_wrapper< primitive_id > >	dependecies ()
	Returns references to all primitive ids on which this primitive depends - inputs, weights, biases, etc.
std::vector< primitive_id >	dependecies () const
	Returns copy of all primitive ids on which this primitive depends - inputs, weights, biases, etc.
	operator primitive_id () const
	Implicit conversion to primiitive id.

Static Public Member Functions
static primitive_type_id	type_id ()

Public Attributes
primitive_id	weights
	Primitive id containing weights data.
primitive_id	bias
	Primitive id containing bias data.
bool	with_activation
	Enable Relu activation.
float	activation_negative_slope
	Relu activation slope.
Public Attributes inherited from cldnn::primitive
const primitive_type_id	type
	Primitive's type id.
const primitive_id	id
	Primitive's id.
fixed_size_vector_ref	input
	List of ids of input primitives.
padding	output_padding
	Requested output padding.

Protected Member Functions
std::vector< std::reference_wrapper< const primitive_id > >	get_dependencies () const override
void	update_dto (dto &dto) const override
Protected Member Functions inherited from cldnn::primitive_base< fully_connected, cldnn_fully_connected_desc >
	primitive_base (const primitive_id &id, const std::vector< primitive_id > &input, const padding &output_padding=padding())
	primitive_base (const cldnn_fully_connected_desc *dto)

Additional Inherited Members
Protected Attributes inherited from cldnn::primitive
primitive_id_arr	_input

Detailed Description

Performs forward fully connected layer (inner product). Also supports built-in Relu cldnn_activation_desc available by setting it in arguments. .

• Equation: Input[F x Y x F] x Output(X) == Weights(B x F x X x F) has to be fulfilled
• Bias has to be linear data [1,1,1,X], where X is equal to number of outputs.

  Format support

  Data type	activation format	weights format
  F32	bfyx	yxfb
  		fyxb
  		bs_xs_xsv8_bsv8
  		bs_x_bsv16
  	yxfb	bfyx
  		yxfb
  		bs_xs_xsv8_bsv8
  F16	bfyx	yxfb
  		fyxb
  		bs_x_bsv16
  	yxfb	bfyx

Definition at line 53 of file fully_connected.hpp.

Constructor & Destructor Documentation

fully_connected()

cldnn::fully_connected::fully_connected ( const primitive_id &  id, const primitive_id &  input, const primitive_id &  weights, const primitive_id &  bias = "", bool  with_activation = false, float  activation_slp = 0.0f, const padding &  output_padding = padding()  )

inline

Constructs fully connected layer.

Parameters

id	This primitive id.
input	Input primitive id.
weights	Primitive id containing weights data.
bias	Primitive id containing bias data. Provide empty string if using Relu without bias.
with_activation	Enable Relu activation.
activation_slp	Relu activation slope.

Definition at line 64 of file fully_connected.hpp.

---

The documentation for this struct was generated from the following file:

• C:/ML/OpenSource/clDNN/api/CPP/fully_connected.hpp