#include <nnet-simple-component.h>

Inheritance diagram for AffineComponent:

Collaboration diagram for AffineComponent:

[legend]

Public Member Functions
virtual int32	InputDim () const
	Returns input-dimension of this component. More...

virtual int32	OutputDim () const
	Returns output-dimension of this component. More...

BaseFloat	OrthonormalConstraint () const

virtual std::string	Info () const
	Returns some text-form information about this component, for diagnostics. More...

virtual void	InitFromConfig (ConfigLine *cfl)
	Initialize, from a ConfigLine object. More...

	AffineComponent ()

virtual std::string	Type () const
	Returns a string such as "SigmoidComponent", describing the type of the object. More...

virtual int32	Properties () const
	Return bitmask of the component's properties. More...

virtual void *	Propagate (const ComponentPrecomputedIndexes indexes, const CuMatrixBase< BaseFloat > &in, CuMatrixBase< BaseFloat > out) const
	Propagate function. More...

virtual void	Backprop (const std::string &debug_info, const ComponentPrecomputedIndexes indexes, const CuMatrixBase< BaseFloat > &in_value, const CuMatrixBase< BaseFloat > &, const CuMatrixBase< BaseFloat > &out_deriv, void memo, Component to_update, CuMatrixBase< BaseFloat > in_deriv) const
	Backprop function; depending on which of the arguments 'to_update' and 'in_deriv' are non-NULL, this can compute input-data derivatives and/or perform model update. More...

virtual void	Read (std::istream &is, bool binary)
	Read function (used after we know the type of the Component); accepts input that is missing the token that describes the component type, in case it has already been consumed. More...

virtual void	Write (std::ostream &os, bool binary) const
	Write component to stream. More...

virtual Component *	Copy () const
	Copies component (deep copy). More...

virtual void	Scale (BaseFloat scale)
	This virtual function when called on – an UpdatableComponent scales the parameters by "scale" when called by an UpdatableComponent. More...

virtual void	Add (BaseFloat alpha, const Component &other)
	This virtual function when called by – an UpdatableComponent adds the parameters of another updatable component, times some constant, to the current parameters. More...

virtual void	PerturbParams (BaseFloat stddev)
	This function is to be used in testing. More...

virtual BaseFloat	DotProduct (const UpdatableComponent &other) const
	Computes dot-product between parameters of two instances of a Component. More...

virtual int32	NumParameters () const
	The following new virtual function returns the total dimension of the parameters in this class. More...

virtual void	Vectorize (VectorBase< BaseFloat > *params) const
	Turns the parameters into vector form. More...

virtual void	UnVectorize (const VectorBase< BaseFloat > &params)
	Converts the parameters from vector form. More...

virtual void	SetParams (const CuVectorBase< BaseFloat > &bias, const CuMatrixBase< BaseFloat > &linear)

const CuVector< BaseFloat > &	BiasParams () const

CuVector< BaseFloat > &	BiasParams ()

const CuMatrix< BaseFloat > &	LinearParams () const

CuMatrix< BaseFloat > &	LinearParams ()

	AffineComponent (const AffineComponent &other)

	AffineComponent (const CuMatrixBase< BaseFloat > &linear_params, const CuVectorBase< BaseFloat > &bias_params, BaseFloat learning_rate)

virtual void	Resize (int32 input_dim, int32 output_dim)

void	Init (int32 input_dim, int32 output_dim, BaseFloat param_stddev, BaseFloat bias_stddev)

Public Member Functions inherited from UpdatableComponent
	UpdatableComponent (const UpdatableComponent &other)

	UpdatableComponent ()

virtual	~UpdatableComponent ()

virtual void	SetUnderlyingLearningRate (BaseFloat lrate)
	Sets the learning rate of gradient descent- gets multiplied by learning_rate_factor_. More...

virtual void	SetActualLearningRate (BaseFloat lrate)
	Sets the learning rate directly, bypassing learning_rate_factor_. More...

virtual void	SetAsGradient ()
	Sets is_gradient_ to true and sets learning_rate_ to 1, ignoring learning_rate_factor_. More...

virtual BaseFloat	LearningRateFactor ()

virtual void	SetLearningRateFactor (BaseFloat lrate_factor)

void	SetUpdatableConfigs (const UpdatableComponent &other)

virtual void	FreezeNaturalGradient (bool freeze)
	freezes/unfreezes NaturalGradient updates, if applicable (to be overriden by components that use Natural Gradient). More...

BaseFloat	LearningRate () const
	Gets the learning rate to be used in gradient descent. More...

BaseFloat	MaxChange () const
	Returns the per-component max-change value, which is interpreted as the maximum change (in l2 norm) in parameters that is allowed per minibatch for this component. More...

void	SetMaxChange (BaseFloat max_change)

BaseFloat	L2Regularization () const
	Returns the l2 regularization constant, which may be set in any updatable component (usually from the config file). More...

void	SetL2Regularization (BaseFloat a)

Public Member Functions inherited from Component
virtual void	StoreStats (const CuMatrixBase< BaseFloat > &in_value, const CuMatrixBase< BaseFloat > &out_value, void *memo)
	This function may store stats on average activation values, and for some component types, the average value of the derivative of the nonlinearity. More...

virtual void	ZeroStats ()
	Components that provide an implementation of StoreStats should also provide an implementation of ZeroStats(), to set those stats to zero. More...

virtual void	GetInputIndexes (const MiscComputationInfo &misc_info, const Index &output_index, std::vector< Index > *desired_indexes) const
	This function only does something interesting for non-simple Components. More...

virtual bool	IsComputable (const MiscComputationInfo &misc_info, const Index &output_index, const IndexSet &input_index_set, std::vector< Index > *used_inputs) const
	This function only does something interesting for non-simple Components, and it exists to make it possible to manage optionally-required inputs. More...

virtual void	ReorderIndexes (std::vector< Index > input_indexes, std::vector< Index > output_indexes) const
	This function only does something interesting for non-simple Components. More...

virtual ComponentPrecomputedIndexes *	PrecomputeIndexes (const MiscComputationInfo &misc_info, const std::vector< Index > &input_indexes, const std::vector< Index > &output_indexes, bool need_backprop) const
	This function must return NULL for simple Components. More...

virtual void	DeleteMemo (void *memo) const
	This virtual function only needs to be overwritten by Components that return a non-NULL memo from their Propagate() function. More...

virtual void	ConsolidateMemory ()
	This virtual function relates to memory management, and avoiding fragmentation. More...

	Component ()

virtual	~Component ()

Protected Member Functions
void	Init (std::string matrix_filename)

virtual void	Update (const std::string &debug_info, const CuMatrixBase< BaseFloat > &in_value, const CuMatrixBase< BaseFloat > &out_deriv)

virtual void	UpdateSimple (const CuMatrixBase< BaseFloat > &in_value, const CuMatrixBase< BaseFloat > &out_deriv)

const AffineComponent &	operator= (const AffineComponent &other)

Protected Member Functions inherited from UpdatableComponent
void	InitLearningRatesFromConfig (ConfigLine *cfl)

std::string	ReadUpdatableCommon (std::istream &is, bool binary)

void	WriteUpdatableCommon (std::ostream &is, bool binary) const

Protected Attributes
CuMatrix< BaseFloat >	linear_params_

CuVector< BaseFloat >	bias_params_

BaseFloat	orthonormal_constraint_

Protected Attributes inherited from UpdatableComponent
BaseFloat	learning_rate_
	learning rate (typically 0.0..0.01) More...

BaseFloat	learning_rate_factor_
	learning rate factor (normally 1.0, but can be set to another < value so that when < you call SetLearningRate(), that value will be scaled by this factor. More...

BaseFloat	l2_regularize_
	L2 regularization constant. More...

bool	is_gradient_
	True if this component is to be treated as a gradient rather than as parameters. More...

BaseFloat	max_change_
	configuration value for imposing max-change More...

Friends
class	NaturalGradientAffineComponent

Additional Inherited Members
Static Public Member Functions inherited from Component
static Component *	ReadNew (std::istream &is, bool binary)
	Read component from stream (works out its type). Dies on error. More...

static Component *	NewComponentOfType (const std::string &type)
	Returns a new Component of the given type e.g. More...

Detailed Description

Definition at line 406 of file nnet-simple-component.h.

Constructor & Destructor Documentation

◆ AffineComponent() [1/3]

AffineComponent ( )

inline

Definition at line 416 of file nnet-simple-component.h.

Referenced by AffineComponent::Copy().

416 : orthonormal_constraint_(0.0) { } // use Init to really initialize.

kaldi::nnet3::AffineComponent::orthonormal_constraint_

BaseFloat orthonormal_constraint_

Definition: nnet-simple-component.h:493

◆ AffineComponent() [2/3]

AffineComponent ( const AffineComponent & other )

explicit

Definition at line 1125 of file nnet-simple-component.cc.

                                                                 :
     UpdatableComponent(component),
     linear_params_(component.linear_params_),
     bias_params_(component.bias_params_),
     orthonormal_constraint_(component.orthonormal_constraint_) { }

◆ AffineComponent() [3/3]

AffineComponent	(	const CuMatrixBase< BaseFloat > &	linear_params,
		const CuVectorBase< BaseFloat > &	bias_params,
		BaseFloat	learning_rate
	)

Definition at line 1131 of file nnet-simple-component.cc.

References CuVectorBase< Real >::Dim(), KALDI_ASSERT, CuMatrixBase< Real >::NumRows(), and UpdatableComponent::SetUnderlyingLearningRate().

                                                          :
     linear_params_(linear_params),
     bias_params_(bias_params),
     orthonormal_constraint_(0.0) {
   SetUnderlyingLearningRate(learning_rate);
   KALDI_ASSERT(linear_params.NumRows() == bias_params.Dim()&&
                bias_params.Dim() != 0);
 }

Member Function Documentation

◆ Add()

void Add	(	BaseFloat	alpha,
		const Component &	other
	)

virtual

This virtual function when called by – an UpdatableComponent adds the parameters of another updatable component, times some constant, to the current parameters.

– a NonlinearComponent (or another component that stores stats, like BatchNormComponent)– it relates to adding stats. Otherwise it will normally do nothing.

Reimplemented from Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1117 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, KALDI_ASSERT, and AffineComponent::linear_params_.

                                                                     {
   const AffineComponent *other =
       dynamic_cast<const AffineComponent*>(&other_in);
   KALDI_ASSERT(other != NULL);
   linear_params_.AddMat(alpha, other->linear_params_);
   bias_params_.AddVec(alpha, other->bias_params_);
 }

◆ Backprop()

void Backprop	(	const std::string &	debug_info,
		const ComponentPrecomputedIndexes *	indexes,
		const CuMatrixBase< BaseFloat > &	in_value,
		const CuMatrixBase< BaseFloat > &	out_value,
		const CuMatrixBase< BaseFloat > &	out_deriv,
		void *	memo,
		Component *	to_update,
		CuMatrixBase< BaseFloat > *	in_deriv
	)		const

virtual

Backprop function; depending on which of the arguments 'to_update' and 'in_deriv' are non-NULL, this can compute input-data derivatives and/or perform model update.

Parameters

[in]	debug_info	The component name, to be printed out in any warning messages.
[in]	indexes	A pointer to some information output by this class's PrecomputeIndexes function (will be NULL for simple components, i.e. those that don't do things like splicing).
[in]	in_value	The matrix that was given as input to the Propagate function. Will be ignored (and may be empty) if Properties()&kBackpropNeedsInput == 0.
[in]	out_value	The matrix that was output from the Propagate function. Will be ignored (and may be empty) if Properties()&kBackpropNeedsOutput == 0
[in]	out_deriv	The derivative at the output of this component.
[in]	memo	This will normally be NULL, but for component types that set the flag kUsesMemo, this will be the return value of the Propagate() function that corresponds to this Backprop() function. Ownership of any pointers is not transferred to the Backprop function; DeleteMemo() will be called to delete it.
[out]	to_update	If model update is desired, the Component to be updated, else NULL. Does not have to be identical to this. If supplied, you can assume that to_update->Properties() & kUpdatableComponent is nonzero.
[out]	in_deriv	The derivative at the input of this component, if needed (else NULL). If Properties()&kBackpropInPlace, may be the same matrix as out_deriv. If Properties()&kBackpropAdds, this is added to by the Backprop routine, else it is set. The component code chooses which mode to work in, based on convenience.

Implements Component.

Definition at line 1260 of file nnet-simple-component.cc.

References CuMatrixBase< Real >::AddMatMat(), UpdatableComponent::is_gradient_, kaldi::kNoTrans, AffineComponent::linear_params_, NVTX_RANGE, AffineComponent::Update(), and AffineComponent::UpdateSimple().

                                                                         {
   NVTX_RANGE("AffineComponent::Backprop");
   AffineComponent *to_update = dynamic_cast<AffineComponent*>(to_update_in);
 
   // Propagate the derivative back to the input.
   // add with coefficient 1.0 since property kBackpropAdds is true.
   // If we wanted to add with coefficient 0.0 we'd need to zero the
   // in_deriv, in case of infinities.
   if (in_deriv)
     in_deriv->AddMatMat(1.0, out_deriv, kNoTrans, linear_params_, kNoTrans,
                         1.0);
 
   if (to_update != NULL) {
     // Next update the model (must do this 2nd so the derivatives we propagate
     // are accurate, in case this == to_update_in.)
     if (to_update->is_gradient_)
       to_update->UpdateSimple(in_value, out_deriv);
     else  // the call below is to a virtual function that may be re-implemented
       to_update->Update(debug_info, in_value, out_deriv);  // by child classes.
   }
 }

◆ BiasParams() [1/2]

const CuVector<BaseFloat>& BiasParams ( ) const

inline

Definition at line 455 of file nnet-simple-component.h.

Referenced by ModelCollapser::CollapseComponentsAffine(), ModelCollapser::CollapseComponentsScale(), SvdApplier::DecomposeComponent(), ModelCollapser::GetDiagonallyPreModifiedComponentIndex(), and kaldi::nnet3::ReduceRankOfComponents().

455 { return bias_params_; }

kaldi::nnet3::AffineComponent::bias_params_

CuVector< BaseFloat > bias_params_

Definition: nnet-simple-component.h:490

◆ BiasParams() [2/2]

CuVector<BaseFloat>& BiasParams ( )

inline

Definition at line 456 of file nnet-simple-component.h.

456 { return bias_params_; }

kaldi::nnet3::AffineComponent::bias_params_

CuVector< BaseFloat > bias_params_

Definition: nnet-simple-component.h:490

◆ Copy()

Component * Copy ( ) const

virtual

Copies component (deep copy).

Implements Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1173 of file nnet-simple-component.cc.

References AffineComponent::AffineComponent().

Referenced by ModelCollapser::CollapseComponentsScale().

                                        {
   AffineComponent *ans = new AffineComponent(*this);
   return ans;
 }

◆ DotProduct()

BaseFloat DotProduct ( const UpdatableComponent & other ) const

virtual

Computes dot-product between parameters of two instances of a Component.

Can be used for computing parameter-norm of an UpdatableComponent.

Implements UpdatableComponent.

Definition at line 1178 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, kaldi::kTrans, AffineComponent::linear_params_, kaldi::TraceMatMat(), and kaldi::VecVec().

                                                                               {
   const AffineComponent *other =
       dynamic_cast<const AffineComponent*>(&other_in);
   return TraceMatMat(linear_params_, other->linear_params_, kTrans)
       + VecVec(bias_params_, other->bias_params_);
 }

◆ Info()

std::string Info ( ) const

virtual

Returns some text-form information about this component, for diagnostics.

Starts with the type of the component. E.g. "SigmoidComponent dim=900", although most components will have much more info.

Reimplemented from UpdatableComponent.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1159 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, kaldi::GetVerboseLevel(), UpdatableComponent::Info(), AffineComponent::linear_params_, AffineComponent::orthonormal_constraint_, and kaldi::nnet3::PrintParameterStats().

Referenced by NaturalGradientAffineComponent::Info().

                                       {
   std::ostringstream stream;
   stream << UpdatableComponent::Info();
   if (orthonormal_constraint_ != 0.0)
     stream << ", orthonormal-constraint=" << orthonormal_constraint_;
   PrintParameterStats(stream, "linear-params", linear_params_,
                       false, // include_mean
                       true, // include_row_norms
                       true, // include_column_norms
                       GetVerboseLevel() >= 2); // include_singular_values
   PrintParameterStats(stream, "bias", bias_params_, true);
   return stream.str();
 }

◆ Init() [1/2]

void Init	(	int32	input_dim,
		int32	output_dim,
		BaseFloat	param_stddev,
		BaseFloat	bias_stddev
	)

Definition at line 1185 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, KALDI_ASSERT, and AffineComponent::linear_params_.

Referenced by ModelCollapser::CollapseComponentsAffine(), and AffineComponent::InitFromConfig().

                                                                           {
   linear_params_.Resize(output_dim, input_dim);
   bias_params_.Resize(output_dim);
   KALDI_ASSERT(output_dim > 0 && input_dim > 0 && param_stddev >= 0.0);
   linear_params_.SetRandn(); // sets to random normally distributed noise.
   linear_params_.Scale(param_stddev);
   bias_params_.SetRandn();
   bias_params_.Scale(bias_stddev);
 }

◆ Init() [2/2]

void Init ( std::string matrix_filename )

protected

Definition at line 1196 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, KALDI_ASSERT, AffineComponent::linear_params_, CuMatrixBase< Real >::NumCols(), CuMatrixBase< Real >::NumRows(), CuMatrixBase< Real >::Range(), and kaldi::ReadKaldiObject().

                                                     {
   CuMatrix<BaseFloat> mat;
   ReadKaldiObject(matrix_filename, &mat); // will abort on failure.
   KALDI_ASSERT(mat.NumCols() >= 2);
   int32 input_dim = mat.NumCols() - 1, output_dim = mat.NumRows();
   linear_params_.Resize(output_dim, input_dim);
   bias_params_.Resize(output_dim);
   linear_params_.CopyFromMat(mat.Range(0, output_dim, 0, input_dim));
   bias_params_.CopyColFromMat(mat, input_dim);
 }

◆ InitFromConfig()

void InitFromConfig ( ConfigLine * cfl )

virtual

Initialize, from a ConfigLine object.

Parameters

[in] cfl A ConfigLine containing any parameters that are needed for initialization. For example: "dim=100 param-stddev=0.1"

Implements Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1207 of file nnet-simple-component.cc.

References ConfigLine::GetValue(), ConfigLine::HasUnusedValues(), AffineComponent::Init(), UpdatableComponent::InitLearningRatesFromConfig(), AffineComponent::InputDim(), KALDI_ASSERT, KALDI_ERR, AffineComponent::orthonormal_constraint_, AffineComponent::OutputDim(), ConfigLine::UnusedValues(), and ConfigLine::WholeLine().

                                                     {
   bool ok = true;
   std::string matrix_filename;
   int32 input_dim = -1, output_dim = -1;
   InitLearningRatesFromConfig(cfl);
   if (cfl->GetValue("matrix", &matrix_filename)) {
     Init(matrix_filename);
     if (cfl->GetValue("input-dim", &input_dim))
       KALDI_ASSERT(input_dim == InputDim() &&
                    "input-dim mismatch vs. matrix.");
     if (cfl->GetValue("output-dim", &output_dim))
       KALDI_ASSERT(output_dim == OutputDim() &&
                    "output-dim mismatch vs. matrix.");
   } else {
     ok = ok && cfl->GetValue("input-dim", &input_dim);
     ok = ok && cfl->GetValue("output-dim", &output_dim);
     BaseFloat param_stddev = 1.0 / std::sqrt(input_dim),
         bias_stddev = 1.0;
     cfl->GetValue("param-stddev", &param_stddev);
     cfl->GetValue("bias-stddev", &bias_stddev);
     Init(input_dim, output_dim,
          param_stddev, bias_stddev);
   }
   cfl->GetValue("orthonormal-constraint", &orthonormal_constraint_);
 
   if (cfl->HasUnusedValues())
     KALDI_ERR << "Could not process these elements in initializer: "
               << cfl->UnusedValues();
   if (!ok)
     KALDI_ERR << "Bad initializer " << cfl->WholeLine();
 }

◆ InputDim()

virtual int32 InputDim ( ) const

inlinevirtual

Returns input-dimension of this component.

Implements Component.

Definition at line 408 of file nnet-simple-component.h.

Referenced by ModelCollapser::CollapseComponentsAffine(), SvdApplier::DecomposeComponent(), SvdApplier::DecomposeComponents(), AffineComponent::InitFromConfig(), NaturalGradientAffineComponent::InitFromConfig(), LinearComponent::InitFromConfig(), AffineComponent::NumParameters(), kaldi::nnet3::ReduceRankOfComponents(), AffineComponent::UnVectorize(), and AffineComponent::Vectorize().

408 { return linear_params_.NumCols(); }

kaldi::nnet3::AffineComponent::linear_params_

CuMatrix< BaseFloat > linear_params_

Definition: nnet-simple-component.h:489

◆ LinearParams() [1/2]

const CuMatrix<BaseFloat>& LinearParams ( ) const

inline

Definition at line 457 of file nnet-simple-component.h.

Referenced by ModelCollapser::CollapseComponentsAffine(), ModelCollapser::CollapseComponentsScale(), kaldi::nnet3::ConstrainOrthonormal(), SvdApplier::DecomposeComponent(), ModelCollapser::GetDiagonallyPreModifiedComponentIndex(), and kaldi::nnet3::ReduceRankOfComponents().

457 { return linear_params_; }

kaldi::nnet3::AffineComponent::linear_params_

CuMatrix< BaseFloat > linear_params_

Definition: nnet-simple-component.h:489

◆ LinearParams() [2/2]

CuMatrix<BaseFloat>& LinearParams ( )

inline

Definition at line 458 of file nnet-simple-component.h.

References PnormComponent::Init().

458 { return linear_params_; }

kaldi::nnet3::AffineComponent::linear_params_

CuMatrix< BaseFloat > linear_params_

Definition: nnet-simple-component.h:489

◆ NumParameters()

int32 NumParameters ( ) const

virtual

The following new virtual function returns the total dimension of the parameters in this class.

Reimplemented from UpdatableComponent.

Definition at line 1323 of file nnet-simple-component.cc.

References AffineComponent::InputDim(), and AffineComponent::OutputDim().

Referenced by AffineComponent::UnVectorize(), and AffineComponent::Vectorize().

                                            {
   return (InputDim() + 1) * OutputDim();
 }

◆ operator=()

const AffineComponent& operator= ( const AffineComponent & other )

protected

◆ OrthonormalConstraint()

BaseFloat OrthonormalConstraint ( ) const

inline

Definition at line 411 of file nnet-simple-component.h.

References Component::Info(), and PnormComponent::InitFromConfig().

Referenced by kaldi::nnet3::ConstrainOrthonormal().

411 { return orthonormal_constraint_; }

kaldi::nnet3::AffineComponent::orthonormal_constraint_

BaseFloat orthonormal_constraint_

Definition: nnet-simple-component.h:493

◆ OutputDim()

virtual int32 OutputDim ( ) const

inlinevirtual

Returns output-dimension of this component.

Implements Component.

Definition at line 409 of file nnet-simple-component.h.

Referenced by ModelCollapser::CollapseComponentsAffine(), ModelCollapser::CollapseComponentsScale(), SvdApplier::DecomposeComponent(), SvdApplier::DecomposeComponents(), AffineComponent::InitFromConfig(), NaturalGradientAffineComponent::InitFromConfig(), LinearComponent::InitFromConfig(), AffineComponent::NumParameters(), kaldi::nnet3::ReduceRankOfComponents(), AffineComponent::UnVectorize(), and AffineComponent::Vectorize().

409 { return linear_params_.NumRows(); }

kaldi::nnet3::AffineComponent::linear_params_

CuMatrix< BaseFloat > linear_params_

Definition: nnet-simple-component.h:489

◆ PerturbParams()

void PerturbParams ( BaseFloat stddev )

virtual

This function is to be used in testing.

It adds unit noise times "stddev" to the parameters of the component.

Implements UpdatableComponent.

Definition at line 1149 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, AffineComponent::linear_params_, CuVectorBase< Real >::SetRandn(), and CuMatrixBase< Real >::SetRandn().

                                                     {
   CuMatrix<BaseFloat> temp_linear_params(linear_params_);
   temp_linear_params.SetRandn();
   linear_params_.AddMat(stddev, temp_linear_params);
 
   CuVector<BaseFloat> temp_bias_params(bias_params_);
   temp_bias_params.SetRandn();
   bias_params_.AddVec(stddev, temp_bias_params);
 }

◆ Propagate()

void * Propagate	(	const ComponentPrecomputedIndexes *	indexes,
		const CuMatrixBase< BaseFloat > &	in,
		CuMatrixBase< BaseFloat > *	out
	)		const

virtual

Propagate function.

Parameters

[in]	indexes	A pointer to some information output by this class's PrecomputeIndexes function (will be NULL for simple components, i.e. those that don't do things like splicing).
[in]	in	The input to this component. Num-columns == InputDim().
[out]	out	The output of this component. Num-columns == OutputDim(). Note: output of this component will be added to the initial value of "out" if Properties()&kPropagateAdds != 0; otherwise the output will be set and the initial value ignored. Each Component chooses whether it is more convenient implementation-wise to add or set, and the calling code has to deal with it.

Returns: Normally returns NULL, but may return a non-NULL value for components which have the flag kUsesMemo set. This value will be passed into the corresponding Backprop routine.

Implements Component.

Definition at line 1242 of file nnet-simple-component.cc.

References CuMatrixBase< Real >::AddMatMat(), AffineComponent::bias_params_, CuMatrixBase< Real >::CopyRowsFromVec(), kaldi::kNoTrans, kaldi::kTrans, and AffineComponent::linear_params_.

                                                                      {
 
   // No need for asserts as they'll happen within the matrix operations.
   out->CopyRowsFromVec(bias_params_); // copies bias_params_ to each row
   // of *out.
   out->AddMatMat(1.0, in, kNoTrans, linear_params_, kTrans, 1.0);
   return NULL;
 }

◆ Properties()

virtual int32 Properties ( ) const

inlinevirtual

Return bitmask of the component's properties.

These properties depend only on the component's type. See enum ComponentProperties.

Implements Component.

Definition at line 418 of file nnet-simple-component.h.

References Component::Add(), PnormComponent::Backprop(), PnormComponent::Copy(), kaldi::nnet3::DotProduct(), kaldi::nnet3::kBackpropAdds, kaldi::nnet3::kBackpropNeedsInput, kaldi::nnet3::kSimpleComponent, kaldi::nnet3::kUpdatableComponent, kaldi::nnet3::NumParameters(), kaldi::nnet3::PerturbParams(), PnormComponent::Propagate(), PnormComponent::Read(), Component::Scale(), and PnormComponent::Write().

                                    {
     return kSimpleComponent|kUpdatableComponent|
         kBackpropNeedsInput|kBackpropAdds;
   }

◆ Read()

void Read	(	std::istream &	is,
		bool	binary
	)

virtual

Read function (used after we know the type of the Component); accepts input that is missing the token that describes the component type, in case it has already been consumed.

Implements Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1289 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, kaldi::nnet3::ExpectToken(), UpdatableComponent::is_gradient_, AffineComponent::linear_params_, AffineComponent::orthonormal_constraint_, kaldi::PeekToken(), kaldi::ReadBasicType(), and UpdatableComponent::ReadUpdatableCommon().

                                                       {
   ReadUpdatableCommon(is, binary);  // read opening tag and learning rate.
   ExpectToken(is, binary, "<LinearParams>");
   linear_params_.Read(is, binary);
   ExpectToken(is, binary, "<BiasParams>");
   bias_params_.Read(is, binary);
   if (PeekToken(is, binary) == 'I') {
     // for back compatibility; we don't write this here any
     // more as it's written and read in Write/ReadUpdatableCommon
     ExpectToken(is, binary, "<IsGradient>");
     ReadBasicType(is, binary, &is_gradient_);
   }
   if (PeekToken(is, binary) == 'O') {
     ExpectToken(is, binary, "<OrthonormalConstraint>");
     ReadBasicType(is, binary, &orthonormal_constraint_);
   } else {
     orthonormal_constraint_ = 0.0;
   }
   ExpectToken(is, binary, "</AffineComponent>");
 }

◆ Resize()

void Resize	(	int32	input_dim,
		int32	output_dim
	)

virtual

Definition at line 1111 of file nnet-simple-component.cc.

References KALDI_ASSERT.

                                                               {
   KALDI_ASSERT(input_dim > 0 && output_dim > 0);
   bias_params_.Resize(output_dim);
   linear_params_.Resize(output_dim, input_dim);
 }

◆ Scale()

void Scale ( BaseFloat scale )

virtual

This virtual function when called on – an UpdatableComponent scales the parameters by "scale" when called by an UpdatableComponent.

– a Nonlinear component (or another component that stores stats, like BatchNormComponent)– it relates to scaling activation stats, not parameters. Otherwise it will normally do nothing.

Reimplemented from Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1100 of file nnet-simple-component.cc.

Referenced by ModelCollapser::GetScaledComponentIndex().

                                            {
   if (scale == 0.0) {
     // If scale == 0.0 we call SetZero() which will get rid of NaN's and inf's.
     linear_params_.SetZero();
     bias_params_.SetZero();
   } else {
     linear_params_.Scale(scale);
     bias_params_.Scale(scale);
   }
 }

◆ SetParams()

void SetParams	(	const CuVectorBase< BaseFloat > &	bias,
		const CuMatrixBase< BaseFloat > &	linear
	)

virtual

Definition at line 1142 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, KALDI_ASSERT, and AffineComponent::linear_params_.

Referenced by ModelCollapser::CollapseComponentsAffine(), ModelCollapser::CollapseComponentsScale(), and kaldi::nnet3::ReduceRankOfComponents().

                                                                        {
   bias_params_ = bias;
   linear_params_ = linear;
   KALDI_ASSERT(bias_params_.Dim() == linear_params_.NumRows());
 }

◆ Type()

virtual std::string Type ( ) const

inlinevirtual

Returns a string such as "SigmoidComponent", describing the type of the object.

Implements Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 417 of file nnet-simple-component.h.

417 { return "AffineComponent"; }

◆ UnVectorize()

void UnVectorize ( const VectorBase< BaseFloat > & params )

virtual

Converts the parameters from vector form.

Reimplemented from UpdatableComponent.

Definition at line 1332 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, VectorBase< Real >::Dim(), AffineComponent::InputDim(), KALDI_ASSERT, AffineComponent::linear_params_, AffineComponent::NumParameters(), AffineComponent::OutputDim(), and VectorBase< Real >::Range().

                                                                      {
   KALDI_ASSERT(params.Dim() == this->NumParameters());
   linear_params_.CopyRowsFromVec(params.Range(0, InputDim() * OutputDim()));
   bias_params_.CopyFromVec(params.Range(InputDim() * OutputDim(),
                                         OutputDim()));
 }

◆ Update()

virtual void Update	(	const std::string &	debug_info,
		const CuMatrixBase< BaseFloat > &	in_value,
		const CuMatrixBase< BaseFloat > &	out_deriv
	)

inlineprotectedvirtual

Reimplemented in NaturalGradientAffineComponent.

Definition at line 476 of file nnet-simple-component.h.

Referenced by AffineComponent::Backprop().

                                                 {
     UpdateSimple(in_value, out_deriv);
   }

◆ UpdateSimple()

void UpdateSimple	(	const CuMatrixBase< BaseFloat > &	in_value,
		const CuMatrixBase< BaseFloat > &	out_deriv
	)

protectedvirtual

Definition at line 1253 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, kaldi::kNoTrans, kaldi::kTrans, UpdatableComponent::learning_rate_, and AffineComponent::linear_params_.

Referenced by AffineComponent::Backprop().

                                                                              {
   bias_params_.AddRowSumMat(learning_rate_, out_deriv, 1.0);
   linear_params_.AddMatMat(learning_rate_, out_deriv, kTrans,
                            in_value, kNoTrans, 1.0);
 }

◆ Vectorize()

void Vectorize ( VectorBase< BaseFloat > * params ) const

virtual

Turns the parameters into vector form.

We put the vector form on the CPU, because in the kinds of situations where we do this, we'll tend to use too much memory for the GPU.

Reimplemented from UpdatableComponent.

Definition at line 1326 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, VectorBase< Real >::Dim(), AffineComponent::InputDim(), KALDI_ASSERT, AffineComponent::linear_params_, AffineComponent::NumParameters(), AffineComponent::OutputDim(), and VectorBase< Real >::Range().

                                                                    {
   KALDI_ASSERT(params->Dim() == this->NumParameters());
   params->Range(0, InputDim() * OutputDim()).CopyRowsFromMat(linear_params_);
   params->Range(InputDim() * OutputDim(),
                 OutputDim()).CopyFromVec(bias_params_);
 }

◆ Write()

void Write	(	std::ostream &	os,
		bool	binary
	)		const

virtual

Write component to stream.

Implements Component.

Reimplemented in NaturalGradientAffineComponent.

Definition at line 1310 of file nnet-simple-component.cc.

References AffineComponent::bias_params_, AffineComponent::linear_params_, AffineComponent::orthonormal_constraint_, kaldi::WriteBasicType(), kaldi::WriteToken(), and UpdatableComponent::WriteUpdatableCommon().

                                                              {
   WriteUpdatableCommon(os, binary);  // Write opening tag and learning rate
   WriteToken(os, binary, "<LinearParams>");
   linear_params_.Write(os, binary);
   WriteToken(os, binary, "<BiasParams>");
   bias_params_.Write(os, binary);
   if (orthonormal_constraint_ != 0.0) {
     WriteToken(os, binary, "<OrthonormalConstraint>");
     WriteBasicType(os, binary, orthonormal_constraint_);
   }
   WriteToken(os, binary, "</AffineComponent>");
 }

Friends And Related Function Documentation

◆ NaturalGradientAffineComponent

friend class NaturalGradientAffineComponent

friend

Definition at line 473 of file nnet-simple-component.h.

Member Data Documentation

◆ bias_params_

◆ linear_params_

◆ orthonormal_constraint_

BaseFloat orthonormal_constraint_

protected

Definition at line 493 of file nnet-simple-component.h.

Referenced by AffineComponent::Info(), LinearComponent::Info(), AffineComponent::InitFromConfig(), NaturalGradientAffineComponent::InitFromConfig(), LinearComponent::InitFromConfig(), AffineComponent::Read(), LinearComponent::Read(), AffineComponent::Write(), NaturalGradientAffineComponent::Write(), and LinearComponent::Write().

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

nnet3/nnet-simple-component.h
nnet3/nnet-simple-component.cc

Public Member Functions

Protected Member Functions

Protected Attributes

Friends

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ AffineComponent() [1/3]

◆ AffineComponent() [2/3]

◆ AffineComponent() [3/3]

Member Function Documentation

◆ Add()

◆ Backprop()

◆ BiasParams() [1/2]

◆ BiasParams() [2/2]

◆ Copy()

◆ DotProduct()

◆ Info()

◆ Init() [1/2]

◆ Init() [2/2]

◆ InitFromConfig()

◆ InputDim()

◆ LinearParams() [1/2]

◆ LinearParams() [2/2]

◆ NumParameters()

◆ operator=()

◆ OrthonormalConstraint()

◆ OutputDim()

◆ PerturbParams()

◆ Propagate()

◆ Properties()

◆ Read()

◆ Resize()

◆ Scale()

◆ SetParams()

◆ Type()

◆ UnVectorize()

◆ Update()

◆ UpdateSimple()

◆ Vectorize()

◆ Write()

Friends And Related Function Documentation

◆ NaturalGradientAffineComponent

Member Data Documentation

◆ bias_params_

◆ linear_params_

◆ orthonormal_constraint_