The ModelBuilder is an abstract class that is overridden by a base class used to programically build new models. More...

Inheritance diagram for MyCaffe.model.ModelBuilder:

Public Types
enum	SCALE_BIAS { NONE , SCALE , BIAS }
	Defines the scale bias type to use. More...

Public Member Functions
	ModelBuilder (string strBaseDir, NetParameter net=null, SolverParameter solver=null)
	The constructor. More...

abstract SolverParameter	CreateSolver ()
	Create the base solver to use. More...

abstract NetParameter	CreateModel (bool bDeploy=false)
	Create the training/testing/deploy model to use. More...

abstract NetParameter	CreateDeployModel ()
	Create the deploy model to use. More...

Protected Member Functions
string	getFileName (string strFile, string strSubDir)
	Returns the full path of the filename using the base directory original set when creating the ModelBuilder. More...

abstract LayerParameter	addExtraLayers (bool bUseBatchNorm=true, double dfLrMult=1.0)
	Add extra layers on top of a 'base' network (e.g. VGGNet or Inception) More...

LayerParameter	findLayer (string strName)
	Find a layer with a given name. More...

NetParameter	createNet (string strName)
	Create the base network parameter for the model and set its name to the 'm_strModel' name. More...

LayerParameter	addDataLayer (string strSource, Phase phase, int nBatchSize=32, bool bOutputLabel=true, TransformationParameter transform=null, string strName="data", bool bSiamese=false)
	Add the Data layer. More...

LayerParameter	addAnnotatedDataLayer (string strSource, Phase phase, int nBatchSize=32, bool bOutputLabel=true, string strLabelMapFile="", SimpleDatum.ANNOTATION_TYPE anno_type=SimpleDatum.ANNOTATION_TYPE.NONE, TransformationParameter transform=null, List< BatchSampler > rgSampler=null)
	Add the Annotated Data layer. More...

List< LayerParameter >	createMultiBoxHead (LayerParameter data, int nNumClasses, List< MultiBoxHeadInfo > rgInfo, List< float > rgPriorVariance, bool bUseObjectness=false, bool bUseBatchNorm=true, double dfLrMult=1.0, bool useScale=true, int nImageHt=0, int nImageWd=0, bool bShareLocation=true, bool bFlip=true, bool bClip=true, double dfOffset=0.5, int nKernelSize=1, int nPad=0, string strConfPostfix="", string strLocPostfix="")
	Create the multi-box head layers. More...

LayerParameter	addConvBNLayer (string strInputLayer, string strOutputLayer, bool bUseBatchNorm, bool bUseRelU, int nNumOutput, int nKernelSize, int nPad, int nStride, double dfLrMult=1.0, int nDilation=1, SCALE_BIAS useScale=SCALE_BIAS.SCALE, string strConvPrefix="", string strConvPostfix="", string strBnPrefix="", string strBnPostfix="_bn", string strScalePrefix="", string strScalePostFix="_scale", string strBiasPrefix="", string strBiasPostfix="_bias", bool bNamedParams=false, string strLayerPostfix="", Phase phaseExclude=Phase.NONE)
	Add convolution, batch-norm layers. More...

LayerParameter	connectAndAddLayer (string fromLayer, LayerParameter toLayer, string fromLayer2=null)
	Connect the from layer to the 'to' layer. More...

LayerParameter	connectAndAddLayer (LayerParameter fromLayer, LayerParameter toLayer, bool bInPlace=false, bool bAdd=true, int nTopIdx=0)
	Connect the from layer to the 'to' layer. More...

LayerParameter	connectAndAddLayer (List< LayerParameter > rgFromLayer, LayerParameter toLayer, bool bAdd=true)
	Connect the from layer to the 'to' layer. More...

LayerParameter	createConvolution (string strName, int nNumOutput, int nKernelSize, int nPad=0, int nStride=1, int nDilation=1)
	Create a new convolution layer parameter. More...

LayerParameter	createPooling (string strName, PoolingParameter.PoolingMethod method, int nKernelSize, int nPad=0, int nStride=1)
	Create a new pooling layer parameter. More...

LayerParameter	addVGGBlock (LayerParameter lastLayer, int nBlockIdx, int nConvIdx, int nNumOutput, int nConvCount, bool? bNoPool, bool bDilatePool=false, int nKernelSize=3, int nPad=1, int nStride=1, int nDilation=1)
	Add a new VGG block. More...

LayerParameter	addVGGfc (LayerParameter lastLayer, int nBlockIdx, int nConvIdx, int nNumOutput, int nDilation, bool bDilated, bool bNoPool, bool bFullConv, bool bReduced, bool bDropout)
	Adds the final layers to the network. More...

LayerParameter	addVGGNetBody (LayerParameter lastLayer, bool bNeedFc=true, bool bFullConv=true, bool bReduced=true, bool bDilated=true, bool bNoPool=false, bool bDropout=false, List< string > rgstrFreezeLayers=null, bool bDilatePool4=false)
	Adds the full VGG body to the network, connecting it to the 'lastLayer'. More...

LayerParameter	addResBody (LayerParameter lastLayer, string strBlockName, int nOut2A, int nOut2B, int nOut2C, int nStride, bool bUseBranch1, int nDilation=1, bool bNamedParams=false, string strLayerPostfix="", Phase phaseExclude=Phase.NONE)
	Adds a ResNet body to the network, connecting it to the 'lastLayer'. More...

void	addExclusion (LayerParameter p, Phase phase)
	Add a phase exclusion. More...

LayerParameter	addResNetBody (string strDataName, int nBlock3Count=4, int nBlock4Count=23, bool bUsePool5=true, bool bUseDilationConv5=false, bool bNamedParams=false, string strLayerPostfix="", Phase phaseExclude=Phase.NONE)
	Create a ResNet101 Body. More...

Protected Attributes
string	m_strBaseDir
	Specifies the base directory that contains the data and models. More...

NetParameter	m_net = new NetParameter()
	Specifies the base net to be altered. More...

SolverParameter	m_solver = new SolverParameter()
	Specifies the base solver to use. More...

Properties
NetParameter	Net `[get]`
	Returns the base net altered by the model builder. More...

SolverParameter	Solver `[get]`
	Returns the base solver. More...

Detailed Description

The ModelBuilder is an abstract class that is overridden by a base class used to programically build new models.

Definition at line 18 of file ModelBuilder.cs.

Member Enumeration Documentation

◆ SCALE_BIAS

enum MyCaffe.model.ModelBuilder.SCALE_BIAS

Defines the scale bias type to use.

Enumerator
NONE	Specifies to not use the scale or bias.
SCALE	Specifies to use the scale.
BIAS	Specifies to use the bias.

Definition at line 36 of file ModelBuilder.cs.

Constructor & Destructor Documentation

◆ ModelBuilder()

MyCaffe.model.ModelBuilder.ModelBuilder	(	string	strBaseDir,
		NetParameter	net = `null`,
		SolverParameter	solver = `null`
	)

The constructor.

Parameters

strBaseDir	Specifies the base directory that contains the data and models.
net	Optionally, specifies the 'base' net parameter that is to be altered (default = null).
solver	Optionally, specifies the 'base' solver parameter to use (default = null).

Definition at line 58 of file ModelBuilder.cs.

Member Function Documentation

◆ addAnnotatedDataLayer()

LayerParameter MyCaffe.model.ModelBuilder.addAnnotatedDataLayer	(	string	strSource,
		Phase	phase,
		int	nBatchSize = `32`,
		bool	bOutputLabel = `true`,
		string	strLabelMapFile = `""`,
		SimpleDatum.ANNOTATION_TYPE	anno_type = `SimpleDatum.ANNOTATION_TYPE.NONE`,
		TransformationParameter	transform = `null`,
		List< BatchSampler >	rgSampler = `null`
	)

protected

Add the Annotated Data layer.

Parameters

strSource	Specifies the data source.
phase	Specifies the phase under which to run the layer (e.g. TRAIN, TEST, RUN).
nBatchSize	Optionally, specifies the batch size (default = 32).
bOutputLabel	Optionally, specifies whether or not to output the label (default = true).
strLabelMapFile	Optionally, specifies the label file (default = "").
anno_type	Optionally, specifies the annotation type (default = NONE).
transform	Optionally, specifies the transformation parameter (default = null, ignored).
rgSampler	Optionally, specifies the list of batch samplers (default = null, ignored).

Returns: The annotated data layer is retunred after it is added to the network.

Definition at line 192 of file ModelBuilder.cs.

◆ addConvBNLayer()

LayerParameter MyCaffe.model.ModelBuilder.addConvBNLayer	(	string	strInputLayer,
		string	strOutputLayer,
		bool	bUseBatchNorm,
		bool	bUseRelU,
		int	nNumOutput,
		int	nKernelSize,
		int	nPad,
		int	nStride,
		double	dfLrMult = `1.0`,
		int	nDilation = `1`,
		SCALE_BIAS	useScale = `SCALE_BIAS.SCALE`,
		string	strConvPrefix = `""`,
		string	strConvPostfix = `""`,
		string	strBnPrefix = `""`,
		string	strBnPostfix = `"_bn"`,
		string	strScalePrefix = `""`,
		string	strScalePostFix = `"_scale"`,
		string	strBiasPrefix = `""`,
		string	strBiasPostfix = `"_bias"`,
		bool	bNamedParams = `false`,
		string	strLayerPostfix = `""`,
		Phase	phaseExclude = `Phase.NONE`
	)

protected

Add convolution, batch-norm layers.

Parameters

strInputLayer	Specifies the input layer.
strOutputLayer	Specifies the output layer.
bUseBatchNorm	Optionally, specifies whether or not to use a batch-norm layer.
bUseRelU	Specifies whether or not to add a RelU layer.
nNumOutput	Specifies the number of output.
nKernelSize	Specifies the kernel size.
nPad	Specifies the pad.
nStride	Specifies the stride.
dfLrMult	Optionally, specifies the default learning rate multiplier (default = 1.0).
nDilation	Optionally, specifies the dilation (default = 1).
useScale	Optionally, specifies whether or not to use a ScaleLayer or BiasLayer (default = SCALE).
strConvPrefix	Optionally, specifies the convolution layer name prefix (default = "").
strConvPostfix	Optionally, specifies the convolution layer name postfix (default = "").
strBnPrefix	Optionally, specifies the batch-norm layer name prefix (default = "").
strBnPostfix	Optionally, specifies the batch-norm layer name postfix (default = "_bn").
strScalePrefix	Optionally, specifies the scale layer name prefix (default = "").
strScalePostFix	Optionally, specifies the scale layer name postfix (default = "_scale")
strBiasPrefix	Optionally, specifies the bias layer name prefix (default = "").
strBiasPostfix	Optionally, specifies the bias layer name postfix (default = "_bias").
bNamedParams	Optionally, specifies to name the parameters (default = false).
strLayerPostfix	Optionally, specifies a layer name postfix (default = "").
phaseExclude	Optionally, specifies a phase to exclude (default = NONE).

Returns: The last layer added is returned.

Definition at line 491 of file ModelBuilder.cs.

◆ addDataLayer()

LayerParameter MyCaffe.model.ModelBuilder.addDataLayer	(	string	strSource,
		Phase	phase,
		int	nBatchSize = `32`,
		bool	bOutputLabel = `true`,
		TransformationParameter	transform = `null`,
		string	strName = `"data"`,
		bool	bSiamese = `false`
	)

protected

Add the Data layer.

Parameters

strSource	Specifies the data source.
phase	Specifies the phase under which to run the layer (e.g. TRAIN, TEST, RUN).
nBatchSize	Optionally, specifies the batch size (default = 32).
bOutputLabel	Optionally, specifies whether or not to output the label (default = true).
transform	Optionally, specifies the transformation parameter (default = null, ignored).
strName	Optionally, specifies a name for the layer and data top name (default = "data").
bSiamese	Optionally, specifies to add data layers for a Siamese net (default = false).

Returns: The Data layer is retunred after it is added to the network.

Definition at line 149 of file ModelBuilder.cs.

◆ addExclusion()

void MyCaffe.model.ModelBuilder.addExclusion	(	LayerParameter	p,
		Phase	phase
	)

protected

Add a phase exclusion.

Parameters

p	Specifies the layer parameter to alter.
phase	Specifies the phase to exclude, or NONE to ignore.

Definition at line 1025 of file ModelBuilder.cs.

◆ addExtraLayers()

abstract LayerParameter MyCaffe.model.ModelBuilder.addExtraLayers	(	bool	bUseBatchNorm = `true`,
		double	dfLrMult = `1.0`
	)

protectedpure virtual

Add extra layers on top of a 'base' network (e.g. VGGNet or Inception)

Parameters

bUseBatchNorm	Optionally, specifies whether or not to use batch normalization layers (default = true).
dfLrMult	Optionally, specifies the learning rate multiplier (default = 1.0).

Implemented in MyCaffe.model.ResNetModelBuilder, MyCaffe.model.ResNetOctConvModelBuilder, and MyCaffe.model.SsdPascalModelBuilder.

◆ addResBody()

LayerParameter MyCaffe.model.ModelBuilder.addResBody	(	LayerParameter	lastLayer,
		string	strBlockName,
		int	nOut2A,
		int	nOut2B,
		int	nOut2C,
		int	nStride,
		bool	bUseBranch1,
		int	nDilation = `1`,
		bool	bNamedParams = `false`,
		string	strLayerPostfix = `""`,
		Phase	phaseExclude = `Phase.NONE`
	)

protected

Adds a ResNet body to the network, connecting it to the 'lastLayer'.

Parameters

lastLayer	Specifies the layer to connect the ResNet net to.
strBlockName	Specifies the block name.
nOut2A	Specifies the output of the first branch.
nOut2B	Specifies the output of the second branch.
nOut2C	Specifies the output of the third branch.
nStride	Specifies the stride used in the first branch.
bUseBranch1	Specifies whether or not to use the first branch.
nDilation	Specifies the dilation used in all branches.
bNamedParams	Optionally, specifies to name the parameters (default = false).
strLayerPostfix	Optionally, specifies a layer name postfix (default = "").
phaseExclude	Optionally, specifies a phase to exclude (default = NONE).

Returns: The last layer added is returned.

Definition at line 959 of file ModelBuilder.cs.

◆ addResNetBody()

LayerParameter MyCaffe.model.ModelBuilder.addResNetBody	(	string	strDataName,
		int	nBlock3Count = `4`,
		int	nBlock4Count = `23`,
		bool	bUsePool5 = `true`,
		bool	bUseDilationConv5 = `false`,
		bool	bNamedParams = `false`,
		string	strLayerPostfix = `""`,
		Phase	phaseExclude = `Phase.NONE`
	)

protected

Create a ResNet101 Body.

Parameters

strDataName	Specifies the last layer name where the ResNet body is to be created from.
nBlock3Count	Specifies the number of blocks in the 3 blocks (default = 4 for ResNet101, use 8 for ResNet152).
nBlock4Count	Speciifes the number of blocks in the 4 blocks (default = 23 for ResNet101, use 36 for ResNet152).
bUsePool5	Specifies whether or not a Pooling layer is to be used as the last layer.
bUseDilationConv5	Specifies whether or not to use dilation on the level 5 block.
bNamedParams	Specifies whether or not to name the parameters (default = false).
strLayerPostfix	Specifies a layer name post-fix (default = "").
phaseExclude	Specifies a phase to exclude, or NONE to ignore (default = NONE).

Returns: The last layer is returned.

Definition at line 1048 of file ModelBuilder.cs.

◆ addVGGBlock()

LayerParameter MyCaffe.model.ModelBuilder.addVGGBlock	(	LayerParameter	lastLayer,
		int	nBlockIdx,
		int	nConvIdx,
		int	nNumOutput,
		int	nConvCount,
		bool?	bNoPool,
		bool	bDilatePool = `false`,
		int	nKernelSize = `3`,
		int	nPad = `1`,
		int	nStride = `1`,
		int	nDilation = `1`
	)

protected

Add a new VGG block.

Parameters

lastLayer	Specifies the last layer that this block is to be connected to.
nBlockIdx	Specifies the block index.
nConvIdx	Specifies the convolution index.
nNumOutput	Specifies the number of outputs.
nConvCount	Specifies the number of convolution layers to add.
bNoPool	When adding the last layer, specifies whether or not to add a pooling (false) or convolution (true) layer. When this parameter is null, the adding of the last layer is skipped.
bDilatePool	Optionally, specifies whether or not to dilate the last pooling layer (default = false).
nKernelSize	Optionally, specifies the kernel size (default = 3).
nPad	Optionally, specifies the pad (default = 1).
nStride	Optionally, specifies the stride (default = 1).
nDilation	Optionally, specifies the dilation (default = 1).

Returns

Definition at line 728 of file ModelBuilder.cs.

◆ addVGGfc()

LayerParameter MyCaffe.model.ModelBuilder.addVGGfc	(	LayerParameter	lastLayer,
		int	nBlockIdx,
		int	nConvIdx,
		int	nNumOutput,
		int	nDilation,
		bool	bDilated,
		bool	bNoPool,
		bool	bFullConv,
		bool	bReduced,
		bool	bDropout
	)

protected

Adds the final layers to the network.

Parameters

lastLayer	Specifies the previous layer to connect the last layers to.
nBlockIdx	Specifies the block index.
nConvIdx	Specifies the convolution index.
nNumOutput	Specifies the number of outputs for the convolution layers.
nDilation	Specifies the dilation to use for the last fully connected convolution layers (used when bFullConv = true).
bDilated	Specifies whether or not dialation is used.
bNoPool	Specifies whether or not pooling is used.
bFullConv	Specifies whether or not full convolution layers are used instead of inner product layers.
bReduced	Specifies whether or not the final layers are used to reduce the data.
bDropout	Specifies whether or not dropout layers are connected.

Returns: The last layer is returned.

Definition at line 777 of file ModelBuilder.cs.

◆ addVGGNetBody()

LayerParameter MyCaffe.model.ModelBuilder.addVGGNetBody	(	LayerParameter	lastLayer,
		bool	bNeedFc = `true`,
		bool	bFullConv = `true`,
		bool	bReduced = `true`,
		bool	bDilated = `true`,
		bool	bNoPool = `false`,
		bool	bDropout = `false`,
		List< string >	rgstrFreezeLayers = `null`,
		bool	bDilatePool4 = `false`
	)

protected

Adds the full VGG body to the network, connecting it to the 'lastLayer'.

Parameters

lastLayer	Specifies the layer to connect the VGG net to.
bNeedFc	Optionally, specifies whether or not to add the fully connected end layers (default = true).
bFullConv	Optionally, specifies whether or not full convolution layers are used instead of inner product layers (default = true).
bReduced	Optionally, specifies whether or not the final layers are used to reduce the data (default = true).
bDilated	Optionally, specifies whether or not dialation is used (default = true).
bNoPool	Optionally, specifies whether or not pooling is used (default = false, use pooling).
bDropout	Optionally, specifies whether or not dropout layers are connected (default = false).
rgstrFreezeLayers	Optionally, specifies a set of layers who's training is to be frozen (default = null to ignore).
bDilatePool4	Optionally, specifies whether or not to dilate pool #4 (default = false).

Returns: The last layer is returned.

Definition at line 915 of file ModelBuilder.cs.

◆ connectAndAddLayer() [1/3]

LayerParameter MyCaffe.model.ModelBuilder.connectAndAddLayer	(	LayerParameter	fromLayer,
		LayerParameter	toLayer,
		bool	bInPlace = `false`,
		bool	bAdd = `true`,
		int	nTopIdx = `0`
	)

protected

Connect the from layer to the 'to' layer.

Parameters

fromLayer	Specifies the layer who's bottom is connected to the toLayer's top.
toLayer	Specifies the layer who's top is connected to the from layer's bottom.
bInPlace	Optionally, specifies whether or not to connect both the top and bottom of the toLayer to the top of the from layer.
bAdd	Optionally, specifies whether or not to add the layer to the network (default = true).
nTopIdx	Optionally, specifies the top index of the item to connect.

Returns: The toLayer is returned as the next layer.

Definition at line 624 of file ModelBuilder.cs.

◆ connectAndAddLayer() [2/3]

LayerParameter MyCaffe.model.ModelBuilder.connectAndAddLayer	(	List< LayerParameter >	rgFromLayer,
		LayerParameter	toLayer,
		bool	bAdd = `true`
	)

protected

Connect the from layer to the 'to' layer.

Parameters

rgFromLayer	Specifies a list of layers who's bottoms are connected to the toLayer's top.
toLayer	Specifies the layer who's top is connected to the from layer's bottom.
bAdd	Optionally, specifies whether or not to add the layer to the network (default = true).

Returns: The toLayer is returned as the next layer.

Definition at line 648 of file ModelBuilder.cs.

◆ connectAndAddLayer() [3/3]

LayerParameter MyCaffe.model.ModelBuilder.connectAndAddLayer	(	string	fromLayer,
		LayerParameter	toLayer,
		string	fromLayer2 = `null`
	)

protected

Connect the from layer to the 'to' layer.

Parameters

fromLayer	Specifies the layer bottom to connect to the toLayer's top.
toLayer	Specifies the layer who's top is connected to the from layer's bottom.
fromLayer2	Optionally, specifies a second input layer.

Returns: The toLayer is returned as the next layer.

Definition at line 602 of file ModelBuilder.cs.

◆ createConvolution()

LayerParameter MyCaffe.model.ModelBuilder.createConvolution	(	string	strName,
		int	nNumOutput,
		int	nKernelSize,
		int	nPad = `0`,
		int	nStride = `1`,
		int	nDilation = `1`
	)

protected

Create a new convolution layer parameter.

Parameters

strName	Specifies the layer name.
nNumOutput	Specifies the number of output.
nKernelSize	Specifies the kernel size.
nPad	Optionally, specifies the pad (default = 0).
nStride	Optionally, specifies the stride (default = 1).
nDilation	Optionally, specifies the dilation (default = 1).

Returns: The convolution layer parameter is returned.

Definition at line 673 of file ModelBuilder.cs.

◆ CreateDeployModel()

abstract NetParameter MyCaffe.model.ModelBuilder.CreateDeployModel ( )

pure virtual

Create the deploy model to use.

Implemented in MyCaffe.model.ResNetModelBuilder, MyCaffe.model.ResNetOctConvModelBuilder, and MyCaffe.model.SsdPascalModelBuilder.

◆ CreateModel()

abstract NetParameter MyCaffe.model.ModelBuilder.CreateModel ( bool bDeploy = false )

pure virtual

Create the training/testing/deploy model to use.

Implemented in MyCaffe.model.ResNetModelBuilder, MyCaffe.model.ResNetOctConvModelBuilder, and MyCaffe.model.SsdPascalModelBuilder.

◆ createMultiBoxHead()

List< LayerParameter > MyCaffe.model.ModelBuilder.createMultiBoxHead	(	LayerParameter	data,
		int	nNumClasses,
		List< MultiBoxHeadInfo >	rgInfo,
		List< float >	rgPriorVariance,
		bool	bUseObjectness = `false`,
		bool	bUseBatchNorm = `true`,
		double	dfLrMult = `1.0`,
		bool	useScale = `true`,
		int	nImageHt = `0`,
		int	nImageWd = `0`,
		bool	bShareLocation = `true`,
		bool	bFlip = `true`,
		bool	bClip = `true`,
		double	dfOffset = `0.5`,
		int	nKernelSize = `1`,
		int	nPad = `0`,
		string	strConfPostfix = `""`,
		string	strLocPostfix = `""`
	)

protected

Create the multi-box head layers.

Parameters

data	Specifies the data layer.
nNumClasses	Specifies the number of classes.
rgInfo	Specifies the info associated with the layers to connect to.
rgPriorVariance	Specifies the prior variance.
bUseObjectness	Optionally, specifies whether or not to use objectness (default = false).
bUseBatchNorm	Optionally, specifies whether or not to use batch-norm layers (default = true).
dfLrMult	Optionally, specifies the learning multiplier (default = 1.0).
useScale	Optionally, specifies whether or not to use scale layers (default = true).
nImageHt	Optionally, specifies the image height (default = 0, ignore).
nImageWd	Optionally, specifies the image width (default = 0, ignore).
bShareLocation	Optionally, specifies whether or not to share the location (default = true).
bFlip	Optionally, specifies whether or not to flip (default = true).
bClip	Optionally, specifies whether or not to clip (default = true).
dfOffset	Optionally, specifies the offset (default = 0.5).
nKernelSize	Optionally, specifies the kernel size (default = 1).
nPad	Optionally, specifies the pad (default = 0).
strConfPostfix	Optionally, specifies the confidence postfix (default = "").
strLocPostfix	Optionally, specifies the location postifix (default = "").

Returns

Definition at line 244 of file ModelBuilder.cs.

◆ createNet()

NetParameter MyCaffe.model.ModelBuilder.createNet ( string strName )

protected

Create the base network parameter for the model and set its name to the 'm_strModel' name.

Parameters

strName Specifies the model name.

Returns: The NetParameter created is returned.

Definition at line 129 of file ModelBuilder.cs.

◆ createPooling()

LayerParameter MyCaffe.model.ModelBuilder.createPooling	(	string	strName,
		PoolingParameter.PoolingMethod	method,
		int	nKernelSize,
		int	nPad = `0`,
		int	nStride = `1`
	)

protected

Create a new pooling layer parameter.

Parameters

strName	Specifies the layer name.
method	Specifies the pooling method.
nKernelSize	Specifies the kernel size.
nPad	Optionally, specifies the pad (default = 0).
nStride	Optionally, specifies the stride (default = 1).

Returns: The pooling layer parameter is returned.

Definition at line 700 of file ModelBuilder.cs.

◆ CreateSolver()

abstract SolverParameter MyCaffe.model.ModelBuilder.CreateSolver ( )

pure virtual

Create the base solver to use.

Returns: The solver parameter created is returned.

Implemented in MyCaffe.model.ResNetModelBuilder, MyCaffe.model.ResNetOctConvModelBuilder, and MyCaffe.model.SsdPascalModelBuilder.

◆ findLayer()

LayerParameter MyCaffe.model.ModelBuilder.findLayer ( string strName )

protected

Find a layer with a given name.

Parameters

strName Specifies the name of the layer to find.

Returns: The layer parameter of the layer with the specified name is returned.

Definition at line 113 of file ModelBuilder.cs.

◆ getFileName()

string MyCaffe.model.ModelBuilder.getFileName	(	string	strFile,
		string	strSubDir
	)

protected

Returns the full path of the filename using the base directory original set when creating the ModelBuilder.

Parameters

strFile	Specifies the partial path of the file.
strSubDir	Specifies the sub-directory off the path (or null for none).

Returns: The full path of the file is returned.

Definition at line 89 of file ModelBuilder.cs.

Member Data Documentation

◆ m_net

NetParameter MyCaffe.model.ModelBuilder.m_net = new NetParameter()

protected

Specifies the base net to be altered.

Definition at line 27 of file ModelBuilder.cs.

◆ m_solver

SolverParameter MyCaffe.model.ModelBuilder.m_solver = new SolverParameter()

protected

Specifies the base solver to use.

Definition at line 31 of file ModelBuilder.cs.

◆ m_strBaseDir

string MyCaffe.model.ModelBuilder.m_strBaseDir

protected

Specifies the base directory that contains the data and models.

Definition at line 23 of file ModelBuilder.cs.

Property Documentation

◆ Net

NetParameter MyCaffe.model.ModelBuilder.Net

get

Returns the base net altered by the model builder.

Definition at line 1105 of file ModelBuilder.cs.

◆ Solver

SolverParameter MyCaffe.model.ModelBuilder.Solver

get

Returns the base solver.

Definition at line 1113 of file ModelBuilder.cs.

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

C:/Data/Data/SS_Projects/Intelligence/GitHub/MyCaffe/MyCaffe.model/ModelBuilder.cs

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Properties

Detailed Description

Member Enumeration Documentation

◆ SCALE_BIAS

Constructor & Destructor Documentation

◆ ModelBuilder()

Member Function Documentation

◆ addAnnotatedDataLayer()

◆ addConvBNLayer()

◆ addDataLayer()

◆ addExclusion()

◆ addExtraLayers()

◆ addResBody()

◆ addResNetBody()

◆ addVGGBlock()

◆ addVGGfc()

◆ addVGGNetBody()

◆ connectAndAddLayer() [1/3]

◆ connectAndAddLayer() [2/3]

◆ connectAndAddLayer() [3/3]

◆ createConvolution()

◆ CreateDeployModel()

◆ CreateModel()

◆ createMultiBoxHead()

◆ createNet()

◆ createPooling()

◆ CreateSolver()

◆ findLayer()

◆ getFileName()

Member Data Documentation

◆ m_net

◆ m_solver

◆ m_strBaseDir

Property Documentation

◆ Net

◆ Solver