TensorRT 10.0.1
nvonnxparser::anonymous_namespace{NvOnnxParser.h} Namespace Reference

Functions

IParsercreateParser (nvinfer1::INetworkDefinition &network, nvinfer1::ILogger &logger)
 Create a new parser object. More...
 
IParserRefittercreateParserRefitter (nvinfer1::IRefitter &refitter, nvinfer1::ILogger &logger)
 Create a new ONNX refitter object. More...
 

Function Documentation

◆ createParser()

IParser * nvonnxparser::anonymous_namespace{NvOnnxParser.h}::createParser ( nvinfer1::INetworkDefinition network,
nvinfer1::ILogger logger 
)
inline

Create a new parser object.

Parameters
networkThe network definition that the parser will write to
loggerThe logger to use
Returns
a new parser object or NULL if an error occurred

Any input dimensions that are constant should not be changed after parsing, because correctness of the translation may rely on those constants. Changing a dynamic input dimension, i.e. one that translates to -1 in TensorRT, to a constant is okay if the constant is consistent with the model. Each instance of the parser is designed to only parse one ONNX model once.

See also
IParser

◆ createParserRefitter()

IParserRefitter * nvonnxparser::anonymous_namespace{NvOnnxParser.h}::createParserRefitter ( nvinfer1::IRefitter refitter,
nvinfer1::ILogger logger 
)
inline

Create a new ONNX refitter object.

Parameters
refitterThe Refitter object used to refit the model
loggerThe logger to use
Returns
a new ParserRefitter object or NULL if an error occurred
See also
IParserRefitter

  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