nvneural::IOnnxGenerationLayer Class Referenceabstract
ONNX export interface for ILayer. More...
#include <nvneural/OnnxTypes.h>
Inheritance diagram for nvneural::IOnnxGenerationLayer:
Public Member Functions | |
| virtual NeuralResult | generateLayerOnnx (IOnnxGenerationHost *pOnnxHost) const noexcept=0 |
| Generates ONNX operators for this layer. 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 = 0x3846c67a6509ae19ul |
| 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
ONNX export interface for ILayer.
Layers should implement this interface to support export to ONNX. By default, layers that do not implement this interface will be replaced by opaque placeholder instances.
Member Function Documentation
◆ generateLayerOnnx()
|
pure virtualnoexcept |
Generates ONNX operators for this layer.
You should not generate operators representing layer post-activation. That is handled by the tool.
If your layer implements multiple fused operations with an activation in between, all of those operations must be emitted by this function. Only the trailing activation is unnecessary in this function.
- Parameters
-
pOnnxHost ONNX host interface. The layer should dump its current state into the graph contained in this object.
- Returns
- NeuralResult::Success normally. Failure if something went wrong and the export should be canceled. You may return NeuralResult::Unsupported to treat this layer as though it had no IOnnxGenerationLayer interface. This is useful if the layer inherits a larger parent class but needs to opt out of ONNX generation. Do not modify the graph if you use this path.
The documentation for this class was generated from the following file:
- Core/Inc/nvneural/OnnxTypes.h
