eVision.EasyDeepLearning.EDeepLearningBenchmark Class Reference

Deep Learning benchmark results. More...

Public Member Functions
None	__init__ (self)
	Constructors.
None	__init__ (self, eVision.EasyDeepLearning.EDeepLearningBenchmark other)
	Constructors.
None	__init__ (self, eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings settings)
	Constructors.
bool	IsValid (self)
	Whether the benchmark is ready and valid.
None	Load (self, str filePath)
	Load a eVision.EasyDeepLearning.EDeepLearningBenchmark object.
None	Load (self, eVision.ESerializer|None serializer)
	Load a eVision.EasyDeepLearning.EDeepLearningBenchmark object.
None	Save (self, str filePath)
	Save the eVision.EasyDeepLearning.EDeepLearningBenchmark object.
None	Save (self, eVision.ESerializer|None serializer)
	Save the eVision.EasyDeepLearning.EDeepLearningBenchmark object.
eVision.EError	ErrorCode (self)
	Error code thrown when computing the benchmark.
str	ErrorDescription (self)
	Description of the error that occurred during the benchmark.
int	NumImagesByInference (self)
	The number of images that were processed together in one inference call to exploit the specified batch size.
int	NumInferences (self)
	Number of inference made in the benchmark.
list[float]	TimePerInference (self)
	Inference time (in nanoseconds) for all the inference calls (an inference processes eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference images together).
int	NumDroppedInference (self)
	The number of dropped inference call.
float	MinLatency (self)
	Minimum latency (time between the acquisition of an image and the computation of its result).
float	MaxLatency (self)
	Maximum latency (time between the acquisition of an image and the computation of its result).
float	MeanLatency (self)
	Mean latency.
float	Throughput (self)
	Number of images per second that were processed by the Deep Learning tool.
float	MinTimePerInference (self)
	Minimum time for a single inference call (in seconds).
float	MaxTimePerInference (self)
	Maximum time for a single inference call (in seconds).
float	MeanTimePerInference (self)
	Mean time for a single inference call (in seconds).
float	StdTimePerInference (self)
	Standard deviation of the inference time for a single inference call (in seconds).
float	MinTimePerImage (self)
	Minimum time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MinTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).
float	MaxTimePerImage (self)
	Maximum time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MaxTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).
float	MeanTimePerImage (self)
	Mean time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MeanTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).
float	StdTimePerImage (self)
	Standard deviation of the time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.StdTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).
int	DatasetWidth (self)
	Width of the dataset used to run the benchmark.
int	DatasetHeight (self)
	Height of the dataset used to run the benchmark.
int	BenchmarkWidth (self)
	Width of the images generated in the benchmark.
int	BenchmarkHeight (self)
	Height of the images generated in the benchmark.
eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings	BenchmarkSettings (self)
	Settings of the benchmark.

Detailed Description

Deep Learning benchmark results.

Remarks:

License(s): EasyClassify or EasySegment or EasyLocate

Constructor & Destructor Documentation

__init__() [1/3]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.__init__

(

self

)

Constructors.

__init__() [2/3]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.__init__	(		self,
		eVision.EasyDeepLearning.EDeepLearningBenchmark	other )

Constructors.

Parameters

other

The other eVision.EasyDeepLearning.EDeepLearningBenchmark that we will copy into this one.

__init__() [3/3]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.__init__	(		self,
		eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings	settings )

Constructors.

Parameters

settings

The settings used to configure this eVision.EasyDeepLearning.EDeepLearningBenchmark.

Member Function Documentation

BenchmarkHeight()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.BenchmarkHeight

(

self

)

Height of the images generated in the benchmark.

It depends on eVision.EasyDeepLearning.EDeepLearningBenchmark.DatasetHeight and eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings.InternalResizeDisabled).

BenchmarkSettings()

eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings eVision.EasyDeepLearning.EDeepLearningBenchmark.BenchmarkSettings

(

self

)

Settings of the benchmark.

BenchmarkWidth()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.BenchmarkWidth

(

self

)

Width of the images generated in the benchmark.

It depends on eVision.EasyDeepLearning.EDeepLearningBenchmark.DatasetWidth and eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings.InternalResizeDisabled).

DatasetHeight()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.DatasetHeight

(

self

)

Height of the dataset used to run the benchmark.

DatasetWidth()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.DatasetWidth

(

self

)

Width of the dataset used to run the benchmark.

ErrorCode()

eVision.EError eVision.EasyDeepLearning.EDeepLearningBenchmark.ErrorCode

(

self

)

Error code thrown when computing the benchmark.

ErrorDescription()

str eVision.EasyDeepLearning.EDeepLearningBenchmark.ErrorDescription

(

self

)

Description of the error that occurred during the benchmark.

IsValid()

bool eVision.EasyDeepLearning.EDeepLearningBenchmark.IsValid

(

self

)

Whether the benchmark is ready and valid.

Load() [1/2]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.Load	(		self,
		eVision.ESerializer | None	serializer )

Load a eVision.EasyDeepLearning.EDeepLearningBenchmark object.

The given eVision.ESerializer must have been created for reading.

Parameters

serializer

Pointer to the eVision.ESerializer created for reading.

Load() [2/2]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.Load	(		self,
		str	filePath )

Load a eVision.EasyDeepLearning.EDeepLearningBenchmark object.

The given eVision.ESerializer must have been created for reading.

Parameters

filePath

File path.

MaxLatency()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MaxLatency

(

self

)

Maximum latency (time between the acquisition of an image and the computation of its result).

MaxTimePerImage()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MaxTimePerImage

(

self

)

Maximum time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MaxTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).

MaxTimePerInference()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MaxTimePerInference

(

self

)

Maximum time for a single inference call (in seconds).

MeanLatency()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MeanLatency

(

self

)

Mean latency.

MeanTimePerImage()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MeanTimePerImage

(

self

)

Mean time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MeanTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).

MeanTimePerInference()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MeanTimePerInference

(

self

)

Mean time for a single inference call (in seconds).

MinLatency()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MinLatency

(

self

)

Minimum latency (time between the acquisition of an image and the computation of its result).

MinTimePerImage()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MinTimePerImage

(

self

)

Minimum time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.MinTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).

MinTimePerInference()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.MinTimePerInference

(

self

)

Minimum time for a single inference call (in seconds).

NumDroppedInference()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.NumDroppedInference

(

self

)

The number of dropped inference call.

Dropped inference call happens when simulating a camera framerate and the processing speed isn't fast enough.

NumImagesByInference()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference

(

self

)

The number of images that were processed together in one inference call to exploit the specified batch size.

For EasyClassify and EasyLocate, this is equal to the batch size. For EasySegment, it depends on the number of patches that are extracted from input images.

NumInferences()

int eVision.EasyDeepLearning.EDeepLearningBenchmark.NumInferences

(

self

)

Number of inference made in the benchmark.

It is equal to eVision.EasyDeepLearning.EDeepLearningBenchmarkSettings.NumInferences

Save() [1/2]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.Save	(		self,
		eVision.ESerializer | None	serializer )

Save the eVision.EasyDeepLearning.EDeepLearningBenchmark object.

The given eVision.ESerializer must have been created for writing.

Parameters

serializer

Pointer to the eVision.ESerializer created for writing.

Save() [2/2]

None eVision.EasyDeepLearning.EDeepLearningBenchmark.Save	(		self,
		str	filePath )

Save the eVision.EasyDeepLearning.EDeepLearningBenchmark object.

The given eVision.ESerializer must have been created for writing.

Parameters

filePath

File path.

StdTimePerImage()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.StdTimePerImage

(

self

)

Standard deviation of the time in seconds for an image (equal to eVision.EasyDeepLearning.EDeepLearningBenchmark.StdTimePerInference / eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference).

StdTimePerInference()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.StdTimePerInference

(

self

)

Standard deviation of the inference time for a single inference call (in seconds).

Throughput()

float eVision.EasyDeepLearning.EDeepLearningBenchmark.Throughput

(

self

)

Number of images per second that were processed by the Deep Learning tool.

TimePerInference()

list[float] eVision.EasyDeepLearning.EDeepLearningBenchmark.TimePerInference

(

self

)

Inference time (in nanoseconds) for all the inference calls (an inference processes eVision.EasyDeepLearning.EDeepLearningBenchmark.NumImagesByInference images together).

A time of 0 means that the images for this inference call were dropped (not processed).