nvneural::IPrototypeLayer Class Reference

IPrototypeLayer is an optional interface that marks a layer as a prototype layer. More...

#include <nvneural/layers/IPrototypeLayer.h>

Inheritance diagram for nvneural::IPrototypeLayer:

Public Member Functions
virtual NeuralResult	addForwardEvalCall (const char *pName, const char *pCall) noexcept=0
	Adds a call script for a particular tensor format and backend. More...
virtual NeuralResult	addParameterKey (const char *pKey) noexcept=0
	Adds a named parameter to the layer. More...
virtual NeuralResult	addPrototypeCode (const char *pName, const char *pCode) noexcept=0
	Adds a kernel definition for a particular tensor format and backend. More...
virtual NeuralResult	loadPrototypeInfo (const IParameterNode *pParameters) noexcept=0
	Loads prototype definitions from a parameter node. More...
virtual NeuralResult	setPrototypeInfo (std::size_t inputsCount, const char *pType, NetworkBackendId backend, TensorFormat format, const char *pDim, const char *pImpl) noexcept=0
	Sets details of the prototype layer explicitly. More...
Public Member Functions inherited from nvneural::IRefObject
virtual RefCount	addRef () const noexcept=0
	Increments the object's reference count. More...
virtual const void *	queryInterface (TypeId interface) const noexcept=0
	This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.
virtual void *	queryInterface (TypeId interface) noexcept=0
	Retrieves a new object interface pointer. More...
virtual RefCount	release () const noexcept=0
	Decrements the object's reference count and destroy the object if the reference count reaches zero. More...

Static Public Attributes
static const IRefObject::TypeId	typeID = 0xb24d9c4a56a0c3f7ul
	Interface TypeId for InterfaceOf purposes.
Static Public Attributes inherited from nvneural::IRefObject
static const TypeId	typeID = 0x14ecc3f9de638e1dul
	Interface TypeId for InterfaceOf purposes.

Additional Inherited Members
Public Types inherited from nvneural::IRefObject
using	RefCount = std::uint32_t
	Typedef used to track the number of active references to an object.
using	TypeId = std::uint64_t
	Every interface must define a unique TypeId. This should be randomized.
Protected Member Functions inherited from nvneural::IRefObject
virtual	~IRefObject ()=default
	A protected destructor prevents accidental stack-allocation of IRefObjects or use with other smart pointer classes like std::unique_ptr.

Detailed Description

IPrototypeLayer is an optional interface that marks a layer as a prototype layer.

Prototype layers carry layer inference code as part of the model definition; this interface allows network builders such as XmlNetworkBuilder to connect the layer to its implementation.

Member Function Documentation

addForwardEvalCall()

virtual NeuralResult nvneural::IPrototypeLayer::addForwardEvalCall ( const char *  pName, const char *  pCall  )

pure virtualnoexcept

Adds a call script for a particular tensor format and backend.

Replaces previous definitions of that kernel.

Parameters

pName	Name of the prototype (e.g., "cuda_fp32_nchw")
pCall	Call script for the prototype

addParameterKey()

virtual NeuralResult nvneural::IPrototypeLayer::addParameterKey ( const char *  pKey )

pure virtualnoexcept

Adds a named parameter to the layer.

Parameters

pKey	Parameter name that should be included in kernel calls

addPrototypeCode()

virtual NeuralResult nvneural::IPrototypeLayer::addPrototypeCode ( const char *  pName, const char *  pCode  )

pure virtualnoexcept

Adds a kernel definition for a particular tensor format and backend.

Replaces previous definitions of that kernel.

Parameters

pName	Name of the prototype provided (e.g., "cuda_fp32_nchw")
pCode	Kernel code for the prototype

loadPrototypeInfo()

virtual NeuralResult nvneural::IPrototypeLayer::loadPrototypeInfo ( const IParameterNode *  pParameters )

pure virtualnoexcept

Loads prototype definitions from a parameter node.

This is not part of an XML <Parameters> declaration; see the definition of XmlNetworkBuilder for how parameters are packed into an IParameterNode for this call.

Network builders use loadPrototypeInfo; generated code will typically use the more fine-grained setPrototypeInfo and addThing functions.

Parameters

pParameters

Attributes of the prototype layer

setPrototypeInfo()

virtual NeuralResult nvneural::IPrototypeLayer::setPrototypeInfo ( std::size_t  inputsCount, const char *  pType, NetworkBackendId  backend, TensorFormat  format, const char *  pDim, const char *  pImpl  )

pure virtualnoexcept

Sets details of the prototype layer explicitly.

Parameters

inputsCount	Number of inputs provided to the layer
pType	Type string to report in ILayer::serializedType
backend	Network backend ID to report in ILayer
format	Tensor format to report in ILayer
pDim	Dimension script string
pImpl	Implementation type string

---

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

• Core/Inc/nvneural/layers/IPrototypeLayer.h