nvneural::INetworkDebugger Class Reference

INetworkDebugger defines a callback interface for network inference. More...

#include <nvneural/NetworkTypes.h>

Inheritance diagram for nvneural::INetworkDebugger:

Public Member Functions
virtual NeuralResult	beginLayer (ILayer *pLayer) noexcept=0
	Indicates that layer inference is beginning. More...
virtual NeuralResult	beginNetwork (INetworkRuntime *pNetworkRuntime, ILayer *pModelLayer, ILayerList *pExternLayerList) noexcept=0
	Indicates the beginning of subnetwork inference within a pass. More...
virtual NeuralResult	beginPass (ILayer **ppInferenceLayers, size_t layerCount) noexcept=0
	Indicates the beginning of an inference pass. More...
virtual NeuralResult	endLayer (ILayer *pLayer) noexcept=0
	Indicates that layer inference has completed. More...
virtual NeuralResult	endNetwork (INetworkRuntime *pNetworkRuntime, ILayer *pModelLayer, ILayerList *pExternLayerList) noexcept=0
	Indicates the end of subnetwork inference within a pass. More...
virtual NeuralResult	endPass () noexcept=0
	Indicates the end of an inference pass.
virtual NeuralResult	skipLayer (ILayer *pLayer) noexcept=0
	Indicates that a layer was optimized out during inference and will be skipped. 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 = 0x44b2d2279e4abca9ul
	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

INetworkDebugger defines a callback interface for network inference.

During inference, each network begins with a beginPass callback and a beginNetwork to initialize the nested network stack. Nested networks (used by network layers) will emit their own beginNetwork and endNetwork pairs.

Individual layers will emit either a beginLayer/endLayer pair or–if bypassed–a single skipLayer call.

The endPass call is emitted at the end of successful inference.

Failure results returned from an INetworkDebugger will abort inference. Inference failures do not unwind debugger state. Error paths should manually reset the debugger if this is necessary.

Member Function Documentation

beginLayer()

virtual NeuralResult nvneural::INetworkDebugger::beginLayer ( ILayer *  pLayer )

pure virtualnoexcept

Indicates that layer inference is beginning.

Parameters

pLayer

Layer being inferenced

beginNetwork()

virtual NeuralResult nvneural::INetworkDebugger::beginNetwork ( INetworkRuntime *  pNetworkRuntime, ILayer *  pModelLayer, ILayerList *  pExternLayerList  )

pure virtualnoexcept

Indicates the beginning of subnetwork inference within a pass.

This callback is also issued at the same time as beginPass for consistency.

Parameters

pNetworkRuntime	The INetworkRuntime performing inference
pModelLayer	The network layer being inferenced; nullptr if this is the outermost network
pExternLayerList	The list of layers being inferenced within this network; nullptr if this is the outermost network

beginPass()

virtual NeuralResult nvneural::INetworkDebugger::beginPass ( ILayer **  ppInferenceLayers, size_t  layerCount  )

pure virtualnoexcept

Indicates the beginning of an inference pass.

Parameters

ppInferenceLayers	Ordered list of layers being inferenced
layerCount	Number of layers in ppInferenceLayers

endLayer()

virtual NeuralResult nvneural::INetworkDebugger::endLayer ( ILayer *  pLayer )

pure virtualnoexcept

Indicates that layer inference has completed.

Parameters

pLayer

The inferenced layer

endNetwork()

virtual NeuralResult nvneural::INetworkDebugger::endNetwork ( INetworkRuntime *  pNetworkRuntime, ILayer *  pModelLayer, ILayerList *  pExternLayerList  )

pure virtualnoexcept

Indicates the end of subnetwork inference within a pass.

This callback is also issued at the same time as endPass for consistency. The arguments match beginNetwork.

Parameters

pNetworkRuntime	The INetworkRuntime performing inference
pModelLayer	The network layer that was inferenced; nullptr if this was the outermost network
pExternLayerList	The list of layers that were inferenced within this network; nullptr if this was the outermost network

skipLayer()

virtual NeuralResult nvneural::INetworkDebugger::skipLayer ( ILayer *  pLayer )

pure virtualnoexcept

Indicates that a layer was optimized out during inference and will be skipped.

Parameters

pLayer

Layer being skipped

---

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

• Core/Inc/nvneural/NetworkTypes.h