EngineConfiguration Class Reference

Inheritance diagram for EngineConfiguration:

Public Member Functions
void	configure (const Properties *fromProperties) override
	Configures the EngineConfiguration from a Properties object throws a ParamException if the properties are not valid.
void	configuration (Properties *toProperties) const override
	Returns the EngineConfiguration parameters as a Properties object.
Properties	toProperties () const
	Convenience functions.
const std::string &	logDomain () const
const std::string &	originalModelPath () const
Public Member Functions inherited from Configurable
virtual void	configureDefaults ()
	Retrieve the properties of this object, replaces values with their defaults and sets it again.
void	registerParameter (ParameterBase *param)
	Register the given Parameter or SubProperty, so that it will be configured during configure()/configuration().
void	unregisterParameter (const ParameterBase *param)
	Remove the given Parameter or SubProperty from the list of registered parameters.
	Configurable (const Configurable &rhs)
	Configurable (Configurable &&rhs) noexcept
Configurable &	operator= (const Configurable &)
Configurable &	operator= (Configurable &&) noexcept

Public Attributes
EngineParameter< std::string >	p_name = {"Name", "", this, ParamRequired::No}
	Type of the engine, i.e.
EngineParameter< std::string >	p_modelFile = {"ModelFile", "", this, ParamRequired::Yes}
	Model file, serialized by the specific framework.
EngineParameter< bool >	p_forceCPU = {"ForceCPU", false, this, ParamRequired::No}
	Whether to force run on CPU even if GPU is present.
EngineParameter< std::vector< std::string > >	p_inputFields
	Input field names expected by the model run by the engine.
EngineParameter< std::vector< std::string > >	p_outputFields
	Output field names returned by the model run by the engine.
EngineParameter< std::vector< std::string > >	p_outputFieldsToIgnore
	Output field names returned by the model run by the engine.
EngineParameter< std::string >	p_version = {"Version", "", this, ParamRequired::No, "EngineVersion"}
	Version of the engine used to train and save the model.
EngineParameter< bool >	p_verbose = {"Verbose", false, this, ParamRequired::No}
	Whether the engine should print a lot of debug information.
Properties	p_engineSpecificParameters
	Parameters that are specific to each engine.
EngineParameter< std::string >	p_yamlConfigPath = {"YamlConfigPath", "", this, ParamRequired::No}
	This is used to determine the path (relative/absolute/resource) of the model file The way to set it is to attach it to the Properties.
Public Attributes inherited from Configurable
Signal	signalParametersChanged
	Emitted whenever one of the registered Parameters' or SubPropertys' signalValueChanged signal was emitted.

Static Public Attributes
static const std::string	p_defaultInputName
static const std::string	p_defaultOutputName

Additional Inherited Members
Protected Attributes inherited from Configurable
std::vector< Param >	m_params
	List of all registered Parameter and SubProperty instances.

Member Function Documentation

configure()

void configure ( const Properties * fromProperties )

overridevirtual

Configures the EngineConfiguration from a Properties object throws a ParamException if the properties are not valid.

Reimplemented from Configurable.

configuration()

void configuration ( Properties * toProperties ) const

overridevirtual

Returns the EngineConfiguration parameters as a Properties object.

Reimplemented from Configurable.

Member Data Documentation

p_name

EngineParameter<std::string> p_name = {"Name", "", this, ParamRequired::No}

Type of the engine, i.e.

"torch", "onnx", "openvino", "tensorrt", ... This string corresponds to the factory name of the specific Engine

p_inputFields

EngineParameter<std::vector<std::string> > p_inputFields

Initial value:

= {"InputFields",
                                                                       std::vector<std::string>{p_defaultInputName},
                                                                       this,
                                                                       ParamRequired::No,
                                                                       "EngineInputFields"}

Input field names expected by the model run by the engine.

p_outputFields

EngineParameter<std::vector<std::string> > p_outputFields

Initial value:

= {"OutputFields",
                                                                        std::vector<std::string>{p_defaultOutputName},
                                                                        this,
                                                                        ParamRequired::No,
                                                                        "EngineOutputFields"}

Output field names returned by the model run by the engine.

p_outputFieldsToIgnore

EngineParameter<std::vector<std::string> > p_outputFieldsToIgnore

Initial value:

= {"OutputFieldsToIgnore",
                                                                                std::vector<std::string>{},
                                                                                this,
                                                                                ParamRequired::No,
                                                                                "EngineOutputFieldsToIgnore"}

Output field names returned by the model run by the engine.

p_engineSpecificParameters

Properties p_engineSpecificParameters

Parameters that are specific to each engine.

These parameters are simply collected from the model yaml configuration and passed over to the selected engine, which is responsible for defining, parsing and using them Example: TorchLibVersion, specific to the torch engine.

p_yamlConfigPath

EngineParameter<std::string> p_yamlConfigPath = {"YamlConfigPath", "", this, ParamRequired::No}

This is used to determine the path (relative/absolute/resource) of the model file The way to set it is to attach it to the Properties.

This is done by the ModelConfiguration. Note that this paramater must be there, otherwise the model file resolution doesn't work.

---

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

• ImFusion/ML/EngineConfiguration.h