TensorRT for RTX 1.6.1
nvonnxparser::IRefitterObserver Class Referenceabstract

Observer interface invoked by IParserRefitter once per refittable engine weight. More...

#include <NvOnnxParser.h>

Public Member Functions

virtual ~IRefitterObserver () noexcept=default
 
virtual void onRefittableWeight (RefitRecord const &record) noexcept=0
 Called once per refittable engine weight as the parser identifies it. More...
 

Detailed Description

Observer interface invoked by IParserRefitter once per refittable engine weight.

Attach via IParserRefitter::setRefitObserver. The intended use is to build a self-describing refit table at build time that can drive nvinfer1::IRefitter::setNamedWeights directly at engine-load time, eliminating the runtime dependency on the original ONNX model structure.

Constructor & Destructor Documentation

◆ ~IRefitterObserver()

virtual nvonnxparser::IRefitterObserver::~IRefitterObserver ( )
virtualdefaultnoexcept

Member Function Documentation

◆ onRefittableWeight()

virtual void nvonnxparser::IRefitterObserver::onRefittableWeight ( RefitRecord const &  record)
pure virtualnoexcept

Called once per refittable engine weight as the parser identifies it.

Invoked during IParserRefitter::refitModelProto, refitFromBytes, or refitFromFile, in the parser's natural traversal order over the ONNX graph. Implementations must not call back into the parser or refitter from this method, and must not retain pointers from record beyond the call.


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

  Copyright © 2024 NVIDIA Corporation
  Privacy Policy | Manage My Privacy | Do Not Sell or Share My Data | Terms of Service | Accessibility | Corporate Policies | Product Security | Contact