The NeuralStyleTransfer object uses the GramLayer, TVLossLayer and LBFGSSolver to perform the neural style transfer algorithm. More...

Inheritance diagram for MyCaffe.extras.NeuralStyleTransfer< T >:

Public Member Functions
	NeuralStyleTransfer (CudaDnn< T > cuda, Log log, CancelEvent evtCancel, string strModelType, string strModel, byte[] rgWeights, bool bCaffeModel, SolverParameter.SolverType solverType=SolverParameter.SolverType.LBFGS, double dfLearningRate=1.5, int nLBFGSCorrections=100, double dfDataScale=1.0, Net< T > netShare=null)
	The constructor. More...

	NeuralStyleTransfer (CudaDnn< T > cuda, Log log, CancelEvent evtCancel, Dictionary< string, Tuple< double, double > > rgLayers, string strModelDesc, byte[] rgWeights, bool bCaffeModel, SolverParameter.SolverType solverType=SolverParameter.SolverType.LBFGS, double dfLearningRate=1.0, int nMaxImageSize=840, int nLBFGSCorrections=100, double dfDataScale=1.0, Net< T > netShare=null)
	The constructor. More...

void	Dispose ()
	Release all resources used. More...

void	SetupHalfSize (bool bAllowHs, bool bAllowOnGram, bool bAllowOnEvent, bool bAllowOnLoss, bool bAllowOnScalar)
	Setup which layers are allowed to use half-sized memory when their convolution counterparts use it. More...

Bitmap	Process (Bitmap bmpStyle, Bitmap bmpContent, int nIterations, int nIntermediateOutput=-1, double dfTvLoss=0, int nMaxSize=-1, bool bEnablePartialSolution=false)
	Process the content image by applying the style to it that was learned from the style image. More...

Bitmap	ProcessNext (out Bitmap bmpIntermediate, out int nIntermediateIdx)
	Process the next partial part of the solution. This function is only valid after calling Process with bEnablePartialSolution = true. More...

Static Public Member Functions
static string	CreateConfigurationString (string strSolver, double dfLearningRate, int nMaxImageSize, int nIterations, int nIntermediateIterations, Dictionary< string, Tuple< double, double > > rgWts, List< int > rgGpuID, int nLBFGSCorrections, double dfDataScale, bool bAllowHs, bool bAllowHsGram, bool bAllowHsEvent, bool bAllowHsScalar, bool bAllowHsLoss)
	The CreateConfigurationString function packs all deep draw settings into a configuration string. More...

static Dictionary< string, Tuple< double, double > >	ParseConfigurationString (string strConfig, out string strSolver, out double dfLearningRate, out int nMaxImageSize, out int nIterations, out int nIntermediateIterations, out List< int > rgGpuID, out int nLBFGSCorrections, out double dfDataScale, out bool bAllowHs, out bool bAllowHsGram, out bool bAllowHsEvent, out bool bAllowHsScalar, out bool bAllowHsLoss)
	The ParseConfigurationString method parses a deep draw configuration string into the actual settings. More...

Events
EventHandler< NeuralStyleIntermediateOutputArgs >	OnIntermediateOutput
	Specifies the event fired after producing intermediate output (e.g. when m_nIntermediateOutput > 0) More...

Detailed Description

The NeuralStyleTransfer object uses the GramLayer, TVLossLayer and LBFGSSolver to perform the neural style transfer algorithm.

See also: minFunc by Marc Schmidt, 2005; ftokarev/caffe-neural-style Github by ftokarev, 2017.; fzliu/style-transfer Github by Frank Liu and Dylan Paiton, 2015.; A Neural Algorithm of Artistic Style by Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge, 2015; Neural Style Transfer: Creating Art with Deep Learning using tf.keras and eager execution by Raymond Yuan, Medium, 2018; Neural Artistic Style Transfer: A Comprehensive Look by Shubhang Desai, Medium, 2017

Definition at line 30 of file NeuralStyleTransfer.cs.

Constructor & Destructor Documentation

◆ NeuralStyleTransfer() [1/2]

MyCaffe.extras.NeuralStyleTransfer< T >.NeuralStyleTransfer	(	CudaDnn< T >	cuda,
		Log	log,
		CancelEvent	evtCancel,
		string	strModelType,
		string	strModel,
		byte[]	rgWeights,
		bool	bCaffeModel,
		SolverParameter.SolverType	solverType = `SolverParameter.SolverType.LBFGS`,
		double	dfLearningRate = `1.5`,
		int	nLBFGSCorrections = `100`,
		double	dfDataScale = `1.0`,
		Net< T >	netShare = `null`
	)

The constructor.

Parameters

cuda	Specifies the instance of CudaDnn to use.
log	Specifies the output log.
evtCancel	Specifies the cancel event used to abort processing.
strModelType	Specifies the model type: 'vgg19', 'vgg16'
strModel	Specifies the network model to use.
rgWeights	Optionally, specifies the weights to use (or null to ignore).
bCaffeModel	Specifies whether or not the weights are in the caffe (true) or mycaffe (false) format.
solverType	Optionally, specifies the solver type to use (default = LBFGS).
dfLearningRate	Optionally, specifies the solver learning rate (default = 1.0).
nLBFGSCorrections	Optionally, specifies the number of LBFGS corrections to use (default = 100, only applies when using LBFGS solver).
dfDataScale	Optionally, specifies the data scaling factor (default = 1.0).
netShare	Optionally, specifies a net to share.

Definition at line 89 of file NeuralStyleTransfer.cs.

◆ NeuralStyleTransfer() [2/2]

MyCaffe.extras.NeuralStyleTransfer< T >.NeuralStyleTransfer	(	CudaDnn< T >	cuda,
		Log	log,
		CancelEvent	evtCancel,
		Dictionary< string, Tuple< double, double > >	rgLayers,
		string	strModelDesc,
		byte[]	rgWeights,
		bool	bCaffeModel,
		SolverParameter.SolverType	solverType = `SolverParameter.SolverType.LBFGS`,
		double	dfLearningRate = `1.0`,
		int	nMaxImageSize = `840`,
		int	nLBFGSCorrections = `100`,
		double	dfDataScale = `1.0`,
		Net< T >	netShare = `null`
	)

The constructor.

Parameters

cuda	Specifies the instance of CudaDnn to use.
log	Specifies the output log.
evtCancel	Specifies the cancel event used to abort processing.
rgLayers	Specifies the layers along with their style and content weights.
strModelDesc	Specifies the network model descriptor to use.
rgWeights	Optionally, specifies the weights to use (or null to ignore).
bCaffeModel	Specifies whether or not the weights are in the caffe (true) or mycaffe (false) format.
solverType	Optionally, specifies the solver type to use (default = LBFGS).
dfLearningRate	Optionally, specifies the solver learning rate (default = 1.0).
nMaxImageSize	Optionally, specifies the default maximum image size (default = 840).
nLBFGSCorrections	Optionally, specifies the LBFGS Corrections (only used when using the LBFGS solver, default = 100).
dfDataScale	Optionally, specifies the data scaling factor (default = 1.0).
netShare	Optionally, specifies a net to share.

Definition at line 137 of file NeuralStyleTransfer.cs.

Member Function Documentation

◆ CreateConfigurationString()

static string MyCaffe.extras.NeuralStyleTransfer< T >.CreateConfigurationString	(	string	strSolver,
		double	dfLearningRate,
		int	nMaxImageSize,
		int	nIterations,
		int	nIntermediateIterations,
		Dictionary< string, Tuple< double, double > >	rgWts,
		List< int >	rgGpuID,
		int	nLBFGSCorrections,
		double	dfDataScale,
		bool	bAllowHs,
		bool	bAllowHsGram,
		bool	bAllowHsEvent,
		bool	bAllowHsScalar,
		bool	bAllowHsLoss
	)

static

The CreateConfigurationString function packs all deep draw settings into a configuration string.

Parameters

strSolver	Specifies the type of solver to use.
dfLearningRate	Specifies the learning rate to use with the solver.
nMaxImageSize	Specifies the maximum image size to use.
nIterations	Specifies the number of iterations to run.
nIntermediateIterations	Specifies how often to output intermediate images if any (a value of 0 disables intermediate output).
rgWts	Specifies the layers to use and their weights for style and content.
rgGpuID	Specifies the GPUIDs on which to run the Neural Style.
nLBFGSCorrections	Specifies the LBFGS corrections to use, only applies when using the LBFGS Solver.
dfDataScale	Specifies the data scale (default = 1.0).
bAllowHs	Specivies to allow half sized memory.
bAllowHsGram	Specifies to allow half sized memory on gram layers.
bAllowHsEvent	Specifies to allow half sized memory on event layers.
bAllowHsScalar	Specifies to allow half sized memory on scalar layers.
bAllowHsLoss	Specifies to allow half sized memory on loss layers.

Returns: The configuration string is returned.

Definition at line 1024 of file NeuralStyleTransfer.cs.

◆ Dispose()

void MyCaffe.extras.NeuralStyleTransfer< T >.Dispose ( )

Release all resources used.

Definition at line 183 of file NeuralStyleTransfer.cs.

◆ ParseConfigurationString()

static Dictionary< string, Tuple< double, double > > MyCaffe.extras.NeuralStyleTransfer< T >.ParseConfigurationString	(	string	strConfig,
		out string	strSolver,
		out double	dfLearningRate,
		out int	nMaxImageSize,
		out int	nIterations,
		out int	nIntermediateIterations,
		out List< int >	rgGpuID,
		out int	nLBFGSCorrections,
		out double	dfDataScale,
		out bool	bAllowHs,
		out bool	bAllowHsGram,
		out bool	bAllowHsEvent,
		out bool	bAllowHsScalar,
		out bool	bAllowHsLoss
	)

static

The ParseConfigurationString method parses a deep draw configuration string into the actual settings.

Parameters

strConfig	Specifies the configuration string to parse.
strSolver	Returns the solver to use.
dfLearningRate	Returns the learning rate to use with the solver.
nMaxImageSize	Returns the maximum image size.
nIterations	Returns the number of iterations to run.
nIntermediateIterations	Returns how often to output intermediate images if any (a value of 0 disables intermediate output).
rgGpuID	Returns the list of GPUIDs on which to run the Neural Style.
nLBFGSCorrections	Returns the LBFGS corrections to use, only applies when using the LBFGS Solver.
dfDataScale	Returns the data scale to use, default = 1.0.
bAllowHs	Returns whether or not half size memory is allowed.
bAllowHsGram	Returns whether or not to allow half size memory in the gram layers.
bAllowHsEvent	Returns whether or not to allow half size memory in the event layers.
bAllowHsScalar	Returns whether or not to allow half size memory in the scalar layers.
bAllowHsLoss	Returns whether or not to allow half size memory in the loss layers.

Returns: Returns a list of layers along with their style and content weights.

Definition at line 1085 of file NeuralStyleTransfer.cs.

◆ Process()

Bitmap MyCaffe.extras.NeuralStyleTransfer< T >.Process	(	Bitmap	bmpStyle,
		Bitmap	bmpContent,
		int	nIterations,
		int	nIntermediateOutput = `-1`,
		double	dfTvLoss = `0`,
		int	nMaxSize = `-1`,
		bool	bEnablePartialSolution = `false`
	)

Process the content image by applying the style to it that was learned from the style image.

Parameters

bmpStyle	Specifies the image used to train the what style to apply to the content.
bmpContent	Specifies the content image to which the style is to be applied.
nIterations	Specifies the number of training iterations.
nIntermediateOutput	Optionally, specifies how often to output an intermediate image.
dfTvLoss	Optionally, specifies the TV-Loss weight for smoothing (default = 0, which disables this loss).
nMaxSize	Optionally, specifies a maximum image size override (default = -1, which uses the default).
bEnablePartialSolution	Optionally, specifies to run the process only up through the next intermediate image. Subsequent calls to ProcessNext moves to the next intermediate image until completed.

Returns: Upon completion the resulting final image is returned, otherwise when using bEnablePartionSolution = true, null is returned.

Definition at line 530 of file NeuralStyleTransfer.cs.

◆ ProcessNext()

Bitmap MyCaffe.extras.NeuralStyleTransfer< T >.ProcessNext	(	out Bitmap	bmpIntermediate,
		out int	nIntermediateIdx
	)

Process the next partial part of the solution. This function is only valid after calling Process with bEnablePartialSolution = true.

Parameters

bmpIntermediate	Returns the intermediate image, if one was created.
nIntermediateIdx	Returns the intermediate index for the image.

Returns: Upon completion, the final Bitmap is returned, otherwise null is returned.

Definition at line 920 of file NeuralStyleTransfer.cs.

◆ SetupHalfSize()

void MyCaffe.extras.NeuralStyleTransfer< T >.SetupHalfSize	(	bool	bAllowHs,
		bool	bAllowOnGram,
		bool	bAllowOnEvent,
		bool	bAllowOnLoss,
		bool	bAllowOnScalar
	)

Setup which layers are allowed to use half-sized memory when their convolution counterparts use it.

Parameters

bAllowHs	Allow half-size memory.
bAllowOnGram	Allow half-size on the gram layers.
bAllowOnEvent	Allow half-size on the event layers.
bAllowOnLoss	Allow half-size on the loss layers.
bAllowOnScalar	Allow half-size on the scalar layers.

Definition at line 215 of file NeuralStyleTransfer.cs.

Event Documentation

◆ OnIntermediateOutput

EventHandler<NeuralStyleIntermediateOutputArgs> MyCaffe.extras.NeuralStyleTransfer< T >.OnIntermediateOutput

Specifies the event fired after producing intermediate output (e.g. when m_nIntermediateOutput > 0)

Definition at line 72 of file NeuralStyleTransfer.cs.

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

C:/Data/Data/SS_Projects/Intelligence/GitHub/MyCaffe/MyCaffe.extras/NeuralStyleTransfer.cs

Public Member Functions

Static Public Member Functions

Events

Detailed Description

Constructor & Destructor Documentation

◆ NeuralStyleTransfer() [1/2]

◆ NeuralStyleTransfer() [2/2]

Member Function Documentation

◆ CreateConfigurationString()

◆ Dispose()

◆ ParseConfigurationString()

◆ Process()

◆ ProcessNext()

◆ SetupHalfSize()

Event Documentation

◆ OnIntermediateOutput