A generalization of MultinomialLogisticLossLayer that takes an "information gain" (infogain) matrix specifying the "value" of all label pairs. More...

#include <infogain_loss_layer.hpp>

Inheritance diagram for caffe::InfogainLossLayer< Dtype >:

Public Member Functions
	InfogainLossLayer (const LayerParameter &param)

virtual void	LayerSetUp (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	Does layer-specific setup: your layer should implement this function as well as Reshape. More...

virtual void	Reshape (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	Adjust the shapes of top blobs and internal buffers to accommodate the shapes of the bottom blobs. More...

virtual int	ExactNumBottomBlobs () const
	Returns the exact number of bottom blobs required by the layer, or -1 if no exact number is required. More...

virtual int	MinBottomBlobs () const
	Returns the minimum number of bottom blobs required by the layer, or -1 if no minimum number is required. More...

virtual int	MaxBottomBlobs () const
	Returns the maximum number of bottom blobs required by the layer, or -1 if no maximum number is required. More...

virtual int	ExactNumTopBlobs () const
	Returns the exact number of top blobs required by the layer, or -1 if no exact number is required. More...

virtual int	MinTopBlobs () const
	Returns the minimum number of top blobs required by the layer, or -1 if no minimum number is required. More...

virtual int	MaxTopBlobs () const
	Returns the maximum number of top blobs required by the layer, or -1 if no maximum number is required. More...

virtual const char *	type () const
	Returns the layer type.

Public Member Functions inherited from caffe::LossLayer< Dtype >
	LossLayer (const LayerParameter &param)

virtual bool	AutoTopBlobs () const
	For convenience and backwards compatibility, instruct the Net to automatically allocate a single top Blob for LossLayers, into which they output their singleton loss, (even if the user didn't specify one in the prototxt, etc.).

virtual bool	AllowForceBackward (const int bottom_index) const

Public Member Functions inherited from caffe::Layer< Dtype >
	Layer (const LayerParameter &param)

void	SetUp (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	Implements common layer setup functionality. More...

Dtype	Forward (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	Given the bottom blobs, compute the top blobs and the loss. More...

void	Backward (const vector< Blob< Dtype > > &top, const vector< bool > &propagate_down, const vector< Blob< Dtype > > &bottom)
	Given the top blob error gradients, compute the bottom blob error gradients. More...

vector< shared_ptr< Blob< Dtype > > > &	blobs ()
	Returns the vector of learnable parameter blobs.

const LayerParameter &	layer_param () const
	Returns the layer parameter.

virtual void	ToProto (LayerParameter *param, bool write_diff=false)
	Writes the layer parameter to a protocol buffer.

Dtype	loss (const int top_index) const
	Returns the scalar loss associated with a top blob at a given index.

void	set_loss (const int top_index, const Dtype value)
	Sets the loss associated with a top blob at a given index.

virtual bool	EqualNumBottomTopBlobs () const
	Returns true if the layer requires an equal number of bottom and top blobs. More...

bool	param_propagate_down (const int param_id)
	Specifies whether the layer should compute gradients w.r.t. a parameter at a particular index given by param_id. More...

void	set_param_propagate_down (const int param_id, const bool value)
	Sets whether the layer should compute gradients w.r.t. a parameter at a particular index given by param_id.

Protected Member Functions
virtual void	Forward_cpu (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	A generalization of MultinomialLogisticLossLayer that takes an "information gain" (infogain) matrix specifying the "value" of all label pairs. More...

virtual void	Backward_cpu (const vector< Blob< Dtype > > &top, const vector< bool > &propagate_down, const vector< Blob< Dtype > > &bottom)
	Computes the infogain loss error gradient w.r.t. the predictions. More...

virtual Dtype	get_normalizer (LossParameter_NormalizationMode normalization_mode, int valid_count)

virtual void	sum_rows_of_H (const Blob< Dtype > *H)
	fill sum_rows_H_ according to matrix H

Protected Member Functions inherited from caffe::Layer< Dtype >
virtual void	Forward_gpu (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)
	Using the GPU device, compute the layer output. Fall back to Forward_cpu() if unavailable.

virtual void	Backward_gpu (const vector< Blob< Dtype > > &top, const vector< bool > &propagate_down, const vector< Blob< Dtype > > &bottom)
	Using the GPU device, compute the gradients for any parameters and for the bottom blobs if propagate_down is true. Fall back to Backward_cpu() if unavailable.

virtual void	CheckBlobCounts (const vector< Blob< Dtype > > &bottom, const vector< Blob< Dtype > > &top)

void	SetLossWeights (const vector< Blob< Dtype > *> &top)

Protected Attributes
shared_ptr< Layer< Dtype > >	softmax_layer_
	The internal SoftmaxLayer used to map predictions to a distribution.

Blob< Dtype >	prob_
	prob stores the output probability predictions from the SoftmaxLayer.

vector< Blob< Dtype > * >	softmax_bottom_vec_
	bottom vector holder used in call to the underlying SoftmaxLayer::Forward

vector< Blob< Dtype > * >	softmax_top_vec_
	top vector holder used in call to the underlying SoftmaxLayer::Forward

Blob< Dtype >	infogain_

Blob< Dtype >	sum_rows_H_

bool	has_ignore_label_
	Whether to ignore instances with a certain label.

int	ignore_label_
	The label indicating that an instance should be ignored.

LossParameter_NormalizationMode	normalization_
	How to normalize the output loss.

int	infogain_axis_

int	outer_num_

int	inner_num_

int	num_labels_

Protected Attributes inherited from caffe::Layer< Dtype >
LayerParameter	layer_param_

Phase	phase_

vector< shared_ptr< Blob< Dtype > > >	blobs_

vector< bool >	param_propagate_down_

vector< Dtype >	loss_

Detailed Description

template<typename Dtype>
class caffe::InfogainLossLayer< Dtype >

A generalization of MultinomialLogisticLossLayer that takes an "information gain" (infogain) matrix specifying the "value" of all label pairs.

Equivalent to the MultinomialLogisticLossLayer if the infogain matrix is the identity.

Parameters

bottom

input Blob vector (length 2-3)

$(N \times C \times H \times W)$ the predictions $\hat{p}$ , a Blob with values in indicating the predicted probability of each of the classes. Each prediction vector $\hat{p}_n$ should sum to 1 as in a probability distribution: $\forall n \sum\limits_{k=1}^K \hat{p}_{nk} = 1$ .
$(N \times 1 \times 1 \times 1)$ the labels , an integer-valued Blob with values $l_n \in [0, 1, 2, ..., K - 1]$ indicating the correct class label among the classes
$(1 \times 1 \times K \times K)$ (optional) the infogain matrix . This must be provided as the third bottom blob input if not provided as the infogain_mat in the InfogainLossParameter. If , this layer is equivalent to the MultinomialLogisticLossLayer.

top

output Blob vector (length 1)

$(1 \times 1 \times 1 \times 1)$ the computed infogain multinomial logistic loss: $E = \frac{-1}{N} \sum\limits_{n=1}^N H_{l_n} \log(\hat{p}_n) = \frac{-1}{N} \sum\limits_{n=1}^N \sum\limits_{k=1}^{K} H_{l_n,k} \log(\hat{p}_{n,k})$ , where $H_{l_n}$ denotes row of .

Member Function Documentation

◆ Backward_cpu()

template<typename Dtype >

void caffe::InfogainLossLayer< Dtype >::Backward_cpu	(	const vector< Blob< Dtype > *> &	top,
		const vector< bool > &	propagate_down,
		const vector< Blob< Dtype > *> &	bottom
	)

protectedvirtual

Computes the infogain loss error gradient w.r.t. the predictions.

Gradients cannot be computed with respect to the label inputs (bottom[1]), so this method ignores bottom[1] and requires !propagate_down[1], crashing if propagate_down[1] is set. (The same applies to the infogain matrix, if provided as bottom[2] rather than in the layer_param.)

Parameters

top	output Blob vector (length 1), providing the error gradient with respect to the outputs $(1 \times 1 \times 1 \times 1)$ This Blob's diff will simply contain the loss_weight* $\lambda$ , as $\lambda$ is the coefficient of this layer's output $\ell_i$ in the overall Net loss $E = \lambda_i \ell_i + \mbox{other loss terms}$ ; hence $\frac{\partial E}{\partial \ell_i} = \lambda_i$ . (*Assuming that this top Blob is not used as a bottom (input) by any other layer of the Net.)
propagate_down	see Layer::Backward. propagate_down[1] must be false as we can't compute gradients with respect to the labels (similarly for propagate_down[2] and the infogain matrix, if provided as bottom[2])
bottom	input Blob vector (length 2-3) $(N \times C \times H \times W)$ the predictions $\hat{p}$ ; Backward computes diff $\frac{\partial E}{\partial \hat{p}}$ $(N \times 1 \times 1 \times 1)$ the labels – ignored as we can't compute their error gradients $(1 \times 1 \times K \times K)$ (optional) the information gain matrix – ignored as its error gradient computation is not implemented.

Implements caffe::Layer< Dtype >.

◆ ExactNumBottomBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::ExactNumBottomBlobs ( ) const

inlinevirtual

Returns the exact number of bottom blobs required by the layer, or -1 if no exact number is required.

This method should be overridden to return a non-negative value if your layer expects some exact number of bottom blobs.

Reimplemented from caffe::LossLayer< Dtype >.

◆ ExactNumTopBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::ExactNumTopBlobs ( ) const

inlinevirtual

Returns the exact number of top blobs required by the layer, or -1 if no exact number is required.

This method should be overridden to return a non-negative value if your layer expects some exact number of top blobs.

Reimplemented from caffe::LossLayer< Dtype >.

◆ Forward_cpu()

template<typename Dtype >

void caffe::InfogainLossLayer< Dtype >::Forward_cpu	(	const vector< Blob< Dtype > *> &	bottom,
		const vector< Blob< Dtype > *> &	top
	)

protectedvirtual

A generalization of MultinomialLogisticLossLayer that takes an "information gain" (infogain) matrix specifying the "value" of all label pairs.

Equivalent to the MultinomialLogisticLossLayer if the infogain matrix is the identity.

Parameters

bottom

input Blob vector (length 2-3)

$(N \times C \times H \times W)$ the predictions $\hat{p}$ , a Blob with values in indicating the predicted probability of each of the classes. Each prediction vector $\hat{p}_n$ should sum to 1 as in a probability distribution: $\forall n \sum\limits_{k=1}^K \hat{p}_{nk} = 1$ .
$(N \times 1 \times 1 \times 1)$ the labels , an integer-valued Blob with values $l_n \in [0, 1, 2, ..., K - 1]$ indicating the correct class label among the classes
$(1 \times 1 \times K \times K)$ (optional) the infogain matrix . This must be provided as the third bottom blob input if not provided as the infogain_mat in the InfogainLossParameter. If , this layer is equivalent to the MultinomialLogisticLossLayer.

top

output Blob vector (length 1)

$(1 \times 1 \times 1 \times 1)$ the computed infogain multinomial logistic loss: $E = \frac{-1}{N} \sum\limits_{n=1}^N H_{l_n} \log(\hat{p}_n) = \frac{-1}{N} \sum\limits_{n=1}^N \sum\limits_{k=1}^{K} H_{l_n,k} \log(\hat{p}_{n,k})$ , where $H_{l_n}$ denotes row of .

Implements caffe::Layer< Dtype >.

◆ get_normalizer()

template<typename Dtype >

Dtype caffe::InfogainLossLayer< Dtype >::get_normalizer	(	LossParameter_NormalizationMode	normalization_mode,
		int	valid_count
	)

protectedvirtual

Read the normalization mode parameter and compute the normalizer based on the blob size. If normalization_mode is VALID, the count of valid outputs will be read from valid_count, unless it is -1 in which case all outputs are assumed to be valid.

◆ LayerSetUp()

template<typename Dtype >

void caffe::InfogainLossLayer< Dtype >::LayerSetUp	(	const vector< Blob< Dtype > *> &	bottom,
		const vector< Blob< Dtype > *> &	top
	)

virtual

Does layer-specific setup: your layer should implement this function as well as Reshape.

Parameters

bottom	the preshaped input blobs, whose data fields store the input data for this layer
top	the allocated but unshaped output blobs

This method should do one-time layer specific setup. This includes reading and processing relevent parameters from the layer_param_. Setting up the shapes of top blobs and internal buffers should be done in Reshape, which will be called before the forward pass to adjust the top blob sizes.

Reimplemented from caffe::LossLayer< Dtype >.

◆ MaxBottomBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::MaxBottomBlobs ( ) const

inlinevirtual

Returns the maximum number of bottom blobs required by the layer, or -1 if no maximum number is required.

This method should be overridden to return a non-negative value if your layer expects some maximum number of bottom blobs.

Reimplemented from caffe::Layer< Dtype >.

◆ MaxTopBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::MaxTopBlobs ( ) const

inlinevirtual

Returns the maximum number of top blobs required by the layer, or -1 if no maximum number is required.

This method should be overridden to return a non-negative value if your layer expects some maximum number of top blobs.

Reimplemented from caffe::Layer< Dtype >.

◆ MinBottomBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::MinBottomBlobs ( ) const

inlinevirtual

Returns the minimum number of bottom blobs required by the layer, or -1 if no minimum number is required.

This method should be overridden to return a non-negative value if your layer expects some minimum number of bottom blobs.

Reimplemented from caffe::Layer< Dtype >.

◆ MinTopBlobs()

template<typename Dtype >

virtual int caffe::InfogainLossLayer< Dtype >::MinTopBlobs ( ) const

inlinevirtual

Returns the minimum number of top blobs required by the layer, or -1 if no minimum number is required.

This method should be overridden to return a non-negative value if your layer expects some minimum number of top blobs.

Reimplemented from caffe::Layer< Dtype >.

◆ Reshape()

template<typename Dtype >

void caffe::InfogainLossLayer< Dtype >::Reshape	(	const vector< Blob< Dtype > *> &	bottom,
		const vector< Blob< Dtype > *> &	top
	)

virtual

Adjust the shapes of top blobs and internal buffers to accommodate the shapes of the bottom blobs.

Parameters

bottom	the input blobs, with the requested input shapes
top	the top blobs, which should be reshaped as needed

This method should reshape top blobs as needed according to the shapes of the bottom (input) blobs, as well as reshaping any internal buffers and making any other necessary adjustments so that the layer can accommodate the bottom blobs.

Reimplemented from caffe::LossLayer< Dtype >.

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

include/caffe/layers/infogain_loss_layer.hpp
src/caffe/layers/infogain_loss_layer.cpp

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

template<typename Dtype> class caffe::InfogainLossLayer< Dtype >

Member Function Documentation

◆ Backward_cpu()

◆ ExactNumBottomBlobs()

◆ ExactNumTopBlobs()

◆ Forward_cpu()

◆ get_normalizer()

◆ LayerSetUp()

◆ MaxBottomBlobs()

◆ MaxTopBlobs()

◆ MinBottomBlobs()

◆ MinTopBlobs()

◆ Reshape()

template<typename Dtype>
class caffe::InfogainLossLayer< Dtype >