#include <nnet-combined-component.h>

Inheritance diagram for LstmNonlinearityComponent:

Collaboration diagram for LstmNonlinearityComponent:

[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...

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...

	LstmNonlinearityComponent ()

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

virtual void	FreezeNaturalGradient (bool freeze)
	virtual More...

	LstmNonlinearityComponent (const LstmNonlinearityComponent &other)

void	Init (int32 cell_dim, bool use_dropout, BaseFloat param_stddev, BaseFloat tanh_self_repair_threshold, BaseFloat sigmoid_self_repair_threshold, BaseFloat self_repair_scale)

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

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)

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	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...

	Component ()

virtual	~Component ()

Private Member Functions
void	InitNaturalGradient ()

const LstmNonlinearityComponent &	operator= (const LstmNonlinearityComponent &other)

Private Attributes
CuMatrix< BaseFloat >	params_

bool	use_dropout_

CuMatrix< double >	value_sum_

CuMatrix< double >	deriv_sum_

CuVector< BaseFloat >	self_repair_config_

CuVector< double >	self_repair_total_

double	count_

OnlineNaturalGradient	preconditioner_

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...

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 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...

Detailed Description

Definition at line 335 of file nnet-combined-component.h.

Constructor & Destructor Documentation

◆ LstmNonlinearityComponent() [1/2]

LstmNonlinearityComponent ( )

inline

Definition at line 342 of file nnet-combined-component.h.

342 : use_dropout_(false) { }

kaldi::nnet3::LstmNonlinearityComponent::use_dropout_

bool use_dropout_

Definition: nnet-combined-component.h:402

◆ LstmNonlinearityComponent() [2/2]

LstmNonlinearityComponent ( const LstmNonlinearityComponent & other )

explicit

Definition at line 1230 of file nnet-combined-component.cc.

                                            :
     UpdatableComponent(other),
     params_(other.params_),
     use_dropout_(other.use_dropout_),
     value_sum_(other.value_sum_),
     deriv_sum_(other.deriv_sum_),
     self_repair_config_(other.self_repair_config_),
     self_repair_total_(other.self_repair_total_),
     count_(other.count_),
     preconditioner_(other.preconditioner_) { }

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.

Definition at line 1128 of file nnet-combined-component.cc.

References LstmNonlinearityComponent::count_, LstmNonlinearityComponent::deriv_sum_, KALDI_ASSERT, LstmNonlinearityComponent::params_, LstmNonlinearityComponent::self_repair_total_, and LstmNonlinearityComponent::value_sum_.

                                                                {
   const LstmNonlinearityComponent *other =
       dynamic_cast<const LstmNonlinearityComponent*>(&other_in);
   KALDI_ASSERT(other != NULL);
   params_.AddMat(alpha, other->params_);
   value_sum_.AddMat(alpha, other->value_sum_);
   deriv_sum_.AddMat(alpha, other->deriv_sum_);
   self_repair_total_.AddVec(alpha, other->self_repair_total_);
   count_ += alpha * other->count_;
 }

◆ 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 1180 of file nnet-combined-component.cc.

References CuVectorBase< Real >::AddColSumMat(), CuVectorBase< Real >::AddVec(), kaldi::cu::BackpropLstmNonlinearity(), LstmNonlinearityComponent::count_, LstmNonlinearityComponent::deriv_sum_, UpdatableComponent::is_gradient_, KALDI_ASSERT, kaldi::kUndefined, UpdatableComponent::learning_rate_, CuMatrixBase< Real >::NumRows(), NVTX_RANGE, LstmNonlinearityComponent::params_, OnlineNaturalGradient::PreconditionDirections(), LstmNonlinearityComponent::preconditioner_, LstmNonlinearityComponent::self_repair_total_, and LstmNonlinearityComponent::value_sum_.

                                              {
   NVTX_RANGE("LstmNonlinearityComponent::Backprop");
 
   if (to_update_in == NULL) {
     cu::BackpropLstmNonlinearity(in_value, params_, out_deriv,
                                  deriv_sum_, self_repair_config_,
                                  count_, in_deriv,
                                  (CuMatrixBase<BaseFloat>*) NULL,
                                  (CuMatrixBase<double>*) NULL,
                                  (CuMatrixBase<double>*) NULL,
                                  (CuMatrixBase<BaseFloat>*) NULL);
   } else {
     LstmNonlinearityComponent *to_update =
         dynamic_cast<LstmNonlinearityComponent*>(to_update_in);
     KALDI_ASSERT(to_update != NULL);
 
     int32 cell_dim = params_.NumCols();
     CuMatrix<BaseFloat> params_deriv(3, cell_dim, kUndefined);
     CuMatrix<BaseFloat> self_repair_total(5, cell_dim, kUndefined);
 
     cu::BackpropLstmNonlinearity(in_value, params_, out_deriv,
                                  deriv_sum_, self_repair_config_,
                                  count_, in_deriv, &params_deriv,
                                  &(to_update->value_sum_),
                                  &(to_update->deriv_sum_),
                                  &self_repair_total);
 
     CuVector<BaseFloat> self_repair_total_sum(5);
     self_repair_total_sum.AddColSumMat(1.0, self_repair_total, 0.0);
     to_update->self_repair_total_.AddVec(1.0, self_repair_total_sum);
     to_update->count_ += static_cast<double>(in_value.NumRows());
 
     BaseFloat scale = 1.0;
     if (!to_update->is_gradient_) {
       to_update->preconditioner_.PreconditionDirections(
           &params_deriv, &scale);
     }
     to_update->params_.AddMat(to_update->learning_rate_ * scale,
                               params_deriv);
   }
 }

◆ ConsolidateMemory()

void ConsolidateMemory ( )

virtual

This virtual function relates to memory management, and avoiding fragmentation.

It is called only once per model, after we do the first minibatch of training. The default implementation does nothing, but it can be overridden by child classes, where it may re-initialize certain quantities that may possibly have been allocated during the forward pass (e.g. certain statistics; OnlineNaturalGradient objects). We use our own CPU-based allocator (see cu-allocator.h) and since it can't do paging since we're not in control of the GPU page table, fragmentation can be a problem. The allocator always tries to put things in 'low-address memory' (i.e. at smaller memory addresses) near the beginning of the block it allocated, to avoid fragmentation; but if permanent things (belonging to the model) are allocated in the forward pass, they can permanently stay in high memory. This function helps to prevent that, by re-allocating those things into low-address memory (It's important that it's called after all the temporary buffers for the forward-backward have been freed, so that there is low-address memory available)).

Reimplemented from Component.

Definition at line 1331 of file nnet-combined-component.cc.

                                                   {
   OnlineNaturalGradient preconditioner_temp(preconditioner_);
   preconditioner_.Swap(&preconditioner_);
 }

◆ Copy()

Component * Copy ( ) const

virtual

Copies component (deep copy).

Implements Component.

Definition at line 1101 of file nnet-combined-component.cc.

                                                  {
   return new LstmNonlinearityComponent(*this);
 }

◆ 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 1146 of file nnet-combined-component.cc.

References KALDI_ASSERT, kaldi::kTrans, LstmNonlinearityComponent::params_, and kaldi::TraceMatMat().

                                               {
   const LstmNonlinearityComponent *other =
       dynamic_cast<const LstmNonlinearityComponent*>(&other_in);
   KALDI_ASSERT(other != NULL);
   return TraceMatMat(params_, other->params_, kTrans);
 }

◆ FreezeNaturalGradient()

void FreezeNaturalGradient ( bool freeze )

virtual

Reimplemented from UpdatableComponent.

Definition at line 1286 of file nnet-combined-component.cc.

References OnlineNaturalGradient::Freeze(), and LstmNonlinearityComponent::preconditioner_.

                                                                  {
   preconditioner_.Freeze(freeze);
 }

◆ 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.

Definition at line 1061 of file nnet-combined-component.cc.

References rnnlm::i, UpdatableComponent::Info(), kaldi::nnet3::PrintParameterStats(), VectorBase< Real >::Scale(), and kaldi::nnet3::SummarizeVector().

                                                 {
   std::ostringstream stream;
   int32 cell_dim = params_.NumCols();
   stream << UpdatableComponent::Info() << ", cell-dim=" << cell_dim
          << ", use-dropout=" << (use_dropout_ ? "true" : "false");
   PrintParameterStats(stream, "w_ic", params_.Row(0));
   PrintParameterStats(stream, "w_fc", params_.Row(1));
   PrintParameterStats(stream, "w_oc", params_.Row(2));
 
   // Note: some of the following code mirrors the code in
   // UpdatableComponent::Info(), in nnet-component-itf.cc.
   if (count_ > 0) {
     stream << ", count=" << std::setprecision(3) << count_
            << std::setprecision(6);
   }
   static const char *nonlin_names[] = { "i_t_sigmoid", "f_t_sigmoid", "c_t_tanh",
                                         "o_t_sigmoid", "m_t_tanh" };
   for (int32 i = 0; i < 5; i++) {
     stream << ", " << nonlin_names[i] << "={";
     stream << " self-repair-lower-threshold=" << self_repair_config_(i)
            << ", self-repair-scale=" << self_repair_config_(i + 5);
 
     if (count_ != 0) {
       BaseFloat self_repaired_proportion =
           self_repair_total_(i) / (count_ * cell_dim);
       stream << ", self-repaired-proportion=" << self_repaired_proportion;
       Vector<double> value_sum(value_sum_.Row(i)),
           deriv_sum(deriv_sum_.Row(i));
       Vector<BaseFloat> value_avg(value_sum), deriv_avg(deriv_sum);
       value_avg.Scale(1.0 / count_);
       deriv_avg.Scale(1.0 / count_);
       stream << ", value-avg=" << SummarizeVector(value_avg)
              << ", deriv-avg=" << SummarizeVector(deriv_avg);
     }
     stream << " }";
   }
   return stream.str();
 }

◆ Init()

void Init	(	int32	cell_dim,
		bool	use_dropout,
		BaseFloat	param_stddev,
		BaseFloat	tanh_self_repair_threshold,
		BaseFloat	sigmoid_self_repair_threshold,
		BaseFloat	self_repair_scale
	)

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

References LstmNonlinearityComponent::count_, LstmNonlinearityComponent::deriv_sum_, LstmNonlinearityComponent::InitNaturalGradient(), KALDI_ASSERT, LstmNonlinearityComponent::params_, CuVector< Real >::Resize(), CuMatrix< Real >::Resize(), LstmNonlinearityComponent::self_repair_config_, LstmNonlinearityComponent::self_repair_total_, LstmNonlinearityComponent::use_dropout_, and LstmNonlinearityComponent::value_sum_.

Referenced by LstmNonlinearityComponent::InitFromConfig().

                                  {
   KALDI_ASSERT(cell_dim > 0 && param_stddev >= 0.0 &&
                tanh_self_repair_threshold >= 0.0 &&
                tanh_self_repair_threshold <= 1.0 &&
                sigmoid_self_repair_threshold >= 0.0 &&
                sigmoid_self_repair_threshold <= 0.25 &&
                self_repair_scale >= 0.0 && self_repair_scale <= 0.1);
   use_dropout_ = use_dropout;
   params_.Resize(3, cell_dim);
   params_.SetRandn();
   params_.Scale(param_stddev);
   value_sum_.Resize(5, cell_dim);
   deriv_sum_.Resize(5, cell_dim);
   self_repair_config_.Resize(10);
   self_repair_config_.Range(0, 5).Set(sigmoid_self_repair_threshold);
   self_repair_config_(2) = tanh_self_repair_threshold;
   self_repair_config_(4) = tanh_self_repair_threshold;
   self_repair_config_.Range(5, 5).Set(self_repair_scale);
   self_repair_total_.Resize(5);
   count_ = 0.0;
   InitNaturalGradient();
 
 }

◆ 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.

Definition at line 1290 of file nnet-combined-component.cc.

References ConfigLine::GetValue(), ConfigLine::HasUnusedValues(), LstmNonlinearityComponent::Init(), UpdatableComponent::InitLearningRatesFromConfig(), KALDI_ERR, LstmNonlinearityComponent::Type(), ConfigLine::UnusedValues(), and ConfigLine::WholeLine().

                                                               {
   InitLearningRatesFromConfig(cfl);
   bool ok = true;
   bool use_dropout = false;
   int32 cell_dim;
   // these self-repair thresholds are the normal defaults for tanh and sigmoid
   // respectively.  If, later on, we decide that we want to support different
   // self-repair config values for the individual sigmoid and tanh
   // nonlinearities, we can modify this code then.
   BaseFloat tanh_self_repair_threshold = 0.2,
       sigmoid_self_repair_threshold = 0.05,
       self_repair_scale = 1.0e-05;
   // param_stddev is the stddev of the parameters.  it may be better to
   // use a smaller value but this was the default in the python scripts
   // for a while.
   BaseFloat param_stddev = 1.0;
   ok = ok && cfl->GetValue("cell-dim", &cell_dim);
   cfl->GetValue("param-stddev", &param_stddev);
   cfl->GetValue("tanh-self-repair-threshold",
                 &tanh_self_repair_threshold);
   cfl->GetValue("sigmoid-self-repair-threshold",
                 &sigmoid_self_repair_threshold);
   cfl->GetValue("self-repair-scale", &self_repair_scale);
   cfl->GetValue("use-dropout", &use_dropout);
 
   // We may later on want to make it possible to initialize the different
   // parameters w_ic, w_fc and w_oc with different biases.  We'll implement
   // that when and if it's needed.
 
   if (cfl->HasUnusedValues())
     KALDI_ERR << "Could not process these elements in initializer: "
               << cfl->UnusedValues();
   if (ok) {
     Init(cell_dim, use_dropout, param_stddev, tanh_self_repair_threshold,
          sigmoid_self_repair_threshold, self_repair_scale);
   } else {
     KALDI_ERR << "Invalid initializer for layer of type "
               << Type() << ": \"" << cfl->WholeLine() << "\"";
   }
 }

◆ InitNaturalGradient()

void InitNaturalGradient ( )

private

Definition at line 1271 of file nnet-combined-component.cc.

References LstmNonlinearityComponent::preconditioner_, OnlineNaturalGradient::SetNumSamplesHistory(), OnlineNaturalGradient::SetRank(), and OnlineNaturalGradient::SetUpdatePeriod().

Referenced by LstmNonlinearityComponent::Init().

                                                     {
   // As regards the configuration for the natural-gradient preconditioner, we
   // don't make it configurable from the command line-- it's unlikely that any
   // differences from changing this would be substantial enough to effectively
   // tune the configuration.  Because the preconditioning code doesn't 'see' the
   // derivatives from individual frames, but only averages over the minibatch,
   // there is a fairly small amount of data available to estimate the Fisher
   // information matrix, so we set the rank, update period and
   // num-samples-history to smaller values than normal.
   preconditioner_.SetRank(20);
   preconditioner_.SetUpdatePeriod(2);
   preconditioner_.SetNumSamplesHistory(1000.0);
 }

◆ InputDim()

int32 InputDim ( ) const

virtual

Returns input-dimension of this component.

Implements Component.

Definition at line 969 of file nnet-combined-component.cc.

Referenced by LstmNonlinearityComponent::ConsolidateMemory().

                                                 {
   int32 cell_dim = value_sum_.NumCols();
   return cell_dim * 5 + (use_dropout_ ? 3 : 0);
 }

◆ 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 1154 of file nnet-combined-component.cc.

Referenced by LstmNonlinearityComponent::ConsolidateMemory().

                                                      {
   return params_.NumRows() * params_.NumCols();
 }

◆ operator=()

const LstmNonlinearityComponent& operator= ( const LstmNonlinearityComponent & other )

private

◆ OutputDim()

int32 OutputDim ( ) const

virtual

Returns output-dimension of this component.

Implements Component.

Definition at line 974 of file nnet-combined-component.cc.

Referenced by LstmNonlinearityComponent::ConsolidateMemory().

                                                  {
   int32 cell_dim = value_sum_.NumCols();
   return cell_dim * 2;
 }

◆ 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 1140 of file nnet-combined-component.cc.

References CuMatrixBase< Real >::SetRandn().

                                                               {
   CuMatrix<BaseFloat> temp_params(params_.NumRows(), params_.NumCols());
   temp_params.SetRandn();
   params_.AddMat(stddev, temp_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 1171 of file nnet-combined-component.cc.

References kaldi::cu::ComputeLstmNonlinearity().

                                         {
   cu::ComputeLstmNonlinearity(in, params_, out);
   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 344 of file nnet-combined-component.h.

                                    {
     return kSimpleComponent|kUpdatableComponent|kBackpropNeedsInput;
   }

◆ 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.

Definition at line 980 of file nnet-combined-component.cc.

References kaldi::nnet3::ExpectToken(), KALDI_ASSERT, kaldi::ReadBasicType(), and kaldi::ReadToken().

                                                                 {
   ReadUpdatableCommon(is, binary);  // Read opening tag and learning rate.
   ExpectToken(is, binary, "<Params>");
   params_.Read(is, binary);
   ExpectToken(is, binary, "<ValueAvg>");
   value_sum_.Read(is, binary);
   ExpectToken(is, binary, "<DerivAvg>");
   deriv_sum_.Read(is, binary);
   ExpectToken(is, binary, "<SelfRepairConfig>");
   self_repair_config_.Read(is, binary);
   ExpectToken(is, binary, "<SelfRepairProb>");
   self_repair_total_.Read(is, binary);
 
   std::string tok;
   ReadToken(is, binary, &tok);
   if (tok == "<UseDropout>") {
     ReadBasicType(is, binary, &use_dropout_);
     ReadToken(is, binary, &tok);
   } else {
     use_dropout_ = false;
   }
   KALDI_ASSERT(tok == "<Count>");
   ReadBasicType(is, binary, &count_);
 
   // For the on-disk format, we normalze value_sum_, deriv_sum_ and
   // self_repair_total_ by dividing by the count, but in memory they are scaled
   // by the count.  [for self_repair_total_, the scaling factor is count_ *
   // cell_dim].
   value_sum_.Scale(count_);
   deriv_sum_.Scale(count_);
   int32 cell_dim = params_.NumCols();
   self_repair_total_.Scale(count_ * cell_dim);
 
   InitNaturalGradient();
 
   ExpectToken(is, binary, "</LstmNonlinearityComponent>");
 
 }

◆ 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.

Definition at line 1112 of file nnet-combined-component.cc.

                                                      {
   if (scale == 0.0) {
     params_.SetZero();
     value_sum_.SetZero();
     deriv_sum_.SetZero();
     self_repair_total_.SetZero();
     count_ = 0.0;
   } else {
     params_.Scale(scale);
     value_sum_.Scale(scale);
     deriv_sum_.Scale(scale);
     self_repair_total_.Scale(scale);
     count_ *= scale;
   }
 }

◆ Type()

virtual std::string Type ( ) const

inlinevirtual

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

Implements Component.

Definition at line 343 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::InitFromConfig().

343 { return "LstmNonlinearityComponent"; }

◆ UnVectorize()

void UnVectorize ( const VectorBase< BaseFloat > & params )

virtual

Converts the parameters from vector form.

Reimplemented from UpdatableComponent.

Definition at line 1164 of file nnet-combined-component.cc.

References VectorBase< Real >::Dim(), KALDI_ASSERT, and kaldi::nnet3::NumParameters().

                                           {
   KALDI_ASSERT(params.Dim() == NumParameters());
   params_.CopyRowsFromVec(params);
 }

◆ 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 1158 of file nnet-combined-component.cc.

References VectorBase< Real >::CopyRowsFromMat(), VectorBase< Real >::Dim(), KALDI_ASSERT, and kaldi::nnet3::NumParameters().

                                                                              {
   KALDI_ASSERT(params->Dim() == NumParameters());
   params->CopyRowsFromMat(params_);
 }

◆ Write()

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

virtual

Write component to stream.

Implements Component.

Definition at line 1019 of file nnet-combined-component.cc.

References MatrixBase< Real >::Scale(), VectorBase< Real >::Scale(), VectorBase< Real >::Write(), MatrixBase< Real >::Write(), kaldi::WriteBasicType(), and kaldi::WriteToken().

                                                                        {
   WriteUpdatableCommon(os, binary);  // Read opening tag and learning rate.
 
   WriteToken(os, binary, "<Params>");
   params_.Write(os, binary);
   WriteToken(os, binary, "<ValueAvg>");
   {
     Matrix<BaseFloat> value_avg(value_sum_);
     if (count_ != 0.0)
       value_avg.Scale(1.0 / count_);
     value_avg.Write(os, binary);
   }
   WriteToken(os, binary, "<DerivAvg>");
   {
     Matrix<BaseFloat> deriv_avg(deriv_sum_);
     if (count_ != 0.0)
       deriv_avg.Scale(1.0 / count_);
     deriv_avg.Write(os, binary);
   }
   WriteToken(os, binary, "<SelfRepairConfig>");
   self_repair_config_.Write(os, binary);
   WriteToken(os, binary, "<SelfRepairProb>");
   {
     int32 cell_dim = params_.NumCols();
     Vector<BaseFloat> self_repair_prob(self_repair_total_);
     if (count_ != 0.0)
       self_repair_prob.Scale(1.0 / (count_ * cell_dim));
     self_repair_prob.Write(os, binary);
   }
   if (use_dropout_) {
     // only write this if true; we have back-compat code in reading anyway.
     // this makes the models without dropout easier to read with older code.
     WriteToken(os, binary, "<UseDropout>");
     WriteBasicType(os, binary, use_dropout_);
   }
   WriteToken(os, binary, "<Count>");
   WriteBasicType(os, binary, count_);
   WriteToken(os, binary, "</LstmNonlinearityComponent>");
 }

◆ ZeroStats()

void ZeroStats ( )

virtual

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

Other components that store other types of statistics (e.g. regarding gradient clipping) should implement ZeroStats() also.

Reimplemented from Component.

Definition at line 1105 of file nnet-combined-component.cc.

                                           {
   value_sum_.SetZero();
   deriv_sum_.SetZero();
   self_repair_total_.SetZero();
   count_ = 0.0;
 }

Member Data Documentation

◆ count_

double count_

private

Definition at line 433 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Add(), LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::ConsolidateMemory(), and LstmNonlinearityComponent::Init().

◆ deriv_sum_

CuMatrix<double> deriv_sum_

private

Definition at line 415 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Add(), LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::ConsolidateMemory(), and LstmNonlinearityComponent::Init().

◆ params_

CuMatrix<BaseFloat> params_

private

Definition at line 398 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Add(), LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::DotProduct(), and LstmNonlinearityComponent::Init().

◆ preconditioner_

OnlineNaturalGradient preconditioner_

private

Definition at line 439 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::ConsolidateMemory(), LstmNonlinearityComponent::FreezeNaturalGradient(), and LstmNonlinearityComponent::InitNaturalGradient().

◆ self_repair_config_

CuVector<BaseFloat> self_repair_config_

private

Definition at line 422 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Init().

◆ self_repair_total_

CuVector<double> self_repair_total_

private

Definition at line 429 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Add(), LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::ConsolidateMemory(), and LstmNonlinearityComponent::Init().

◆ use_dropout_

bool use_dropout_

private

Definition at line 402 of file nnet-combined-component.h.

Referenced by LstmNonlinearityComponent::Init().

◆ value_sum_

CuMatrix<double> value_sum_

private

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

Referenced by LstmNonlinearityComponent::Add(), LstmNonlinearityComponent::Backprop(), LstmNonlinearityComponent::ConsolidateMemory(), and LstmNonlinearityComponent::Init().

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

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

Public Member Functions

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ LstmNonlinearityComponent() [1/2]

◆ LstmNonlinearityComponent() [2/2]

Member Function Documentation

◆ Add()

◆ Backprop()

◆ ConsolidateMemory()

◆ Copy()

◆ DotProduct()

◆ FreezeNaturalGradient()

◆ Info()

◆ Init()

◆ InitFromConfig()

◆ InitNaturalGradient()

◆ InputDim()

◆ NumParameters()

◆ operator=()

◆ OutputDim()

◆ PerturbParams()

◆ Propagate()

◆ Properties()

◆ Read()

◆ Scale()

◆ Type()

◆ UnVectorize()

◆ Vectorize()

◆ Write()

◆ ZeroStats()

Member Data Documentation

◆ count_

◆ deriv_sum_

◆ params_

◆ preconditioner_

◆ self_repair_config_

◆ self_repair_total_

◆ use_dropout_

◆ value_sum_