mmind::dl::MMindInferEngine Class Reference

Defines the infer engine. More...

#include <MMindInferEngine.h>

Public Member Functions
	MMindInferEngine ()
	Constructs the infer engine.
StatusCode	setBatchSize (const unsigned int batchSize, const unsigned int moduleIdx=0)
	Sets the batch size of the model package.
StatusCode	setBatchSize (const std::vector< unsigned int > &batchSize)
	Sets the batch size of the model package.
StatusCode	setFloatPrecision (const FloatPrecisionType floatPrecisionType, const unsigned int moduleIdx=0)
	Sets the float precision of the model package.
StatusCode	setFloatPrecision (const std::vector< FloatPrecisionType > &floatPrecisionType)
	Sets the float precision of the model package.
StatusCode	setDeviceId (const unsigned int deviceId)
	Sets the device ID.
StatusCode	setInferDeviceType (const InferDeviceType type)
	Sets the infer device type.
StatusCode	create (const std::string &modelPath)
	Creates an infer engine for model package inference.
StatusCode	load ()
	Loads the model into memory.
StatusCode	infer (const std::vector< MMindImage > &images)
	Makes image inference using the model package inference engine.
StatusCode	getResults (std::vector< MMindResult > &results)
	Gets the model inference result.
StatusCode	resultVisualization (std::vector< MMindImage > &images)
	Draws all the model results onto the images.
StatusCode	moduleResultVisualization (std::vector< MMindImage > &images, const unsigned int moduleIdx)
	Draws the model results of the specified index onto the images.
std::vector< DeepLearningAlgoType >	getDeepLearningAlgoTypes () const
	Gets the model type list.
void	release ()
	Releases the memory of the model package inference engine.
	MMindInferEngine (const MMindInferEngine &rhs)=delete
	MMindInferEngine (MMindInferEngine &&rhs)
MMindInferEngine &	operator= (const MMindInferEngine &rhs)=delete
MMindInferEngine &	operator= (MMindInferEngine &&rhs)
	~MMindInferEngine ()

Detailed Description

Defines the infer engine.

Constructor & Destructor Documentation

MMindInferEngine() [1/3]

mmind::dl::MMindInferEngine::MMindInferEngine

(

)

Constructs the infer engine.

MMindInferEngine() [2/3]

mmind::dl::MMindInferEngine::MMindInferEngine ( const MMindInferEngine &  rhs )

delete

MMindInferEngine() [3/3]

mmind::dl::MMindInferEngine::MMindInferEngine

(

MMindInferEngine &&

rhs

)

~MMindInferEngine()

mmind::dl::MMindInferEngine::~MMindInferEngine

(

)

Member Function Documentation

create()

StatusCode mmind::dl::MMindInferEngine::create

(

const std::string &

modelPath

)

Creates an infer engine for model package inference.

Parameters

[in]

modelPath

The path to the model package exported from Mech-DLK.

Returns

See StatusCode for details.

getDeepLearningAlgoTypes()

std::vector< DeepLearningAlgoType > mmind::dl::MMindInferEngine::getDeepLearningAlgoTypes

(

)

const

Gets the model type list.

Returns

See DeepLearningAlgoType for details.

getResults()

StatusCode mmind::dl::MMindInferEngine::getResults

(

std::vector< MMindResult > &

results

)

Gets the model inference result.

Parameters

[in]

results

See MMindResult for details.

Returns

See StatusCode for details.

infer()

StatusCode mmind::dl::MMindInferEngine::infer

(

const std::vector< MMindImage > &

images

)

Makes image inference using the model package inference engine.

Parameters

[in]

images

See MMindImage for details.

Returns

See StatusCode for details.

load()

StatusCode mmind::dl::MMindInferEngine::load

(

)

Loads the model into memory.

Returns

See StatusCode for details.

Note

When the type of the infer device is GpuOptimization, it may take 1-5 minutes to optimize the model package.

moduleResultVisualization()

StatusCode mmind::dl::MMindInferEngine::moduleResultVisualization	(	std::vector< MMindImage > &	images,
		const unsigned int	moduleIdx
	)

Draws the model results of the specified index onto the images.

Parameters

[in]	MMindImage	See MMindImage for details.
[in]	moduleIdx	Specified module index in the model package.

Returns

See StatusCode for details.

operator=() [1/2]

MMindInferEngine & mmind::dl::MMindInferEngine::operator= ( const MMindInferEngine &  rhs )

delete

operator=() [2/2]

MMindInferEngine & mmind::dl::MMindInferEngine::operator=

(

MMindInferEngine &&

rhs

)

release()

void mmind::dl::MMindInferEngine::release

(

)

Releases the memory of the model package inference engine.

resultVisualization()

StatusCode mmind::dl::MMindInferEngine::resultVisualization

(

std::vector< MMindImage > &

images

)

Draws all the model results onto the images.

Parameters

[in]

MMindImage

See MMindImage for details.

Returns

See StatusCode for details.

setBatchSize() [1/2]

StatusCode mmind::dl::MMindInferEngine::setBatchSize

(

const std::vector< unsigned int > &

batchSize

)

Sets the batch size of the model package.

Parameters

[in]

batchSize

The batch size of the model package.

Returns

See StatusCode for details.

Note

This function sets the batch size of all modules once.

setBatchSize() [2/2]

StatusCode mmind::dl::MMindInferEngine::setBatchSize	(	const unsigned int	batchSize,
		const unsigned int	moduleIdx = 0
	)

Sets the batch size of the model package.

Parameters

[in]	batchSize	The batch size of the model package.
[in]	moduleIdx	Specified module index in the model package.

Returns

See StatusCode for details.

Note

This function sets only the batch size corresponding to the module index.

setDeviceId()

StatusCode mmind::dl::MMindInferEngine::setDeviceId

(

const unsigned int

deviceId

)

Sets the device ID.

Parameters

[in]

deviceId

The index of the specified GPU during model inference.

Returns

See StatusCode for details.

Note

When the InferDeviceType is set to CPU, the setting of deviceId is invalid.

setFloatPrecision() [1/2]

StatusCode mmind::dl::MMindInferEngine::setFloatPrecision	(	const FloatPrecisionType	floatPrecisionType,
		const unsigned int	moduleIdx = 0
	)

Sets the float precision of the model package.

Parameters

[in]	floatPrecisionType	The float precision of the model package, See FloatPrecisionType for details.
[in]	moduleIdx	Specified module index in the model package.

Returns

See StatusCode for details.

Note

This function sets only the float precision corresponding to the module index.

setFloatPrecision() [2/2]

StatusCode mmind::dl::MMindInferEngine::setFloatPrecision

(

const std::vector< FloatPrecisionType > &

floatPrecisionType

)

Sets the float precision of the model package.

Parameters

[in]

floatPrecisionType

The float precision of the model package, See FloatPrecisionType for details.

Returns

See StatusCode for details.

Note

This function sets the float precision of all modules once.

setInferDeviceType()

StatusCode mmind::dl::MMindInferEngine::setInferDeviceType

(

const InferDeviceType

type

)

Sets the infer device type.

Parameters

[in]

type

See InferDeviceType for details.

Returns

See StatusCode for details.

Note

In CPU mode, the deployment computer must have an INTEL CPU, and in GpuDefault or GpuOptimization mode, the deployment computer must have an NVIDIA GPU.

---

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

• MMindInferEngine.h