#include <nnet-general-component.h>

Inheritance diagram for StatisticsPoolingComponent:

Collaboration diagram for StatisticsPoolingComponent:

Public Member Functions
	StatisticsPoolingComponent ()

	StatisticsPoolingComponent (const StatisticsPoolingComponent &other)

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

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

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

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 > &out_value, const CuMatrixBase< BaseFloat > &out_deriv, void memo, Component , 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	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...

void	SetRequireDirectInput (bool b)

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

Private Member Functions
void	Check () const

StatisticsPoolingComponent &	operator= (const StatisticsPoolingComponent &other)

Private Attributes
int32	input_dim_

int32	input_period_

int32	left_context_

int32	right_context_

int32	num_log_count_features_

bool	output_stddevs_

BaseFloat	variance_floor_

bool	require_direct_input_

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 337 of file nnet-general-component.h.

Constructor & Destructor Documentation

◆ StatisticsPoolingComponent() [1/2]

StatisticsPoolingComponent ( )

Definition at line 575 of file nnet-general-component.cc.

                                                       :
     input_dim_(-1), input_period_(1), left_context_(-1), right_context_(-1),
     num_log_count_features_(0), output_stddevs_(false),
     variance_floor_(1.0e-10), require_direct_input_(false) { }

◆ StatisticsPoolingComponent() [2/2]

StatisticsPoolingComponent ( const StatisticsPoolingComponent & other )

Definition at line 581 of file nnet-general-component.cc.

References StatisticsPoolingComponent::Check().

                                             :
     input_dim_(other.input_dim_), input_period_(other.input_period_),
     left_context_(other.left_context_), right_context_(other.right_context_),
     num_log_count_features_(other.num_log_count_features_),
     output_stddevs_(other.output_stddevs_),
     variance_floor_(other.variance_floor_),
     require_direct_input_(other.require_direct_input_) {
   Check();
 }

Member Function Documentation

◆ 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 826 of file nnet-general-component.cc.

References CuMatrixBase< Real >::AddRowRanges(), CuVectorBase< Real >::ApplyExp(), StatisticsPoolingComponentPrecomputedIndexes::backward_indexes, CuMatrixBase< Real >::ColRange(), CuVectorBase< Real >::CopyColFromMat(), CuVectorBase< Real >::Data(), CuMatrixBase< Real >::DivRowsVec(), StatisticsPoolingComponentPrecomputedIndexes::forward_indexes, StatisticsPoolingComponent::input_dim_, KALDI_ASSERT, kaldi::kUndefined, StatisticsPoolingComponent::num_log_count_features_, CuMatrixBase< Real >::NumRows(), NVTX_RANGE, StatisticsPoolingComponent::output_stddevs_, and CuVectorBase< Real >::SetZero().

                                              {
   NVTX_RANGE("StatisticsPoolingComponent::Backprop");
   KALDI_ASSERT(indexes_in != NULL);
   const StatisticsPoolingComponentPrecomputedIndexes *indexes =
       dynamic_cast<const StatisticsPoolingComponentPrecomputedIndexes*>(
           indexes_in);
   int32 num_rows_out = out_deriv_in.NumRows();
   CuMatrix<BaseFloat> out_deriv(out_deriv_in);
   if (output_stddevs_) {
     // for now we actually ignore the covariance flooring in the backprop- this
     // is an approximation.  Typically the derivatives computed will be quite
     // tiny for floored variances (they should be zero), so it won't affect the
     // derivatives much.
     int32 feature_dim = (input_dim_ - 1) / 2;
     CuSubMatrix<BaseFloat> mean_deriv(out_deriv, 0, num_rows_out,
                                       num_log_count_features_, feature_dim),
         variance_deriv(out_deriv, 0, num_rows_out,
                        num_log_count_features_ + feature_dim, feature_dim),
         mean_value(out_value, 0, num_rows_out,
                    num_log_count_features_, feature_dim),
         stddev_value(out_value, 0, num_rows_out,
                      num_log_count_features_ + feature_dim, feature_dim);
     // we currently have the deriv w.r.t. the stddev.  step 1 is to get it
     // w.r.t. the centered variance.  If the centered variance is s,
     // and the stddev is sqrt(s), then d/ds sqrt(s) = 0.5 / sqrt(s),
     // so we need to multiply variance_deriv by 0.5 / the stddev.
     variance_deriv.DivElements(stddev_value);
     variance_deriv.Scale(0.5);
 
     // the deriv w.r.t. the uncentered variance is the same as w.r.t.  the
     // uncentered variance (since they difer by a constant term of -(mean *
     // mean), but we need to add to dF/dmean, the value -2.0 * mean *
     // dF/dvariance.
     mean_deriv.AddMatMatElements(-2.0, mean_value, variance_deriv, 1.0);
   }
   // now we have to account for the effect of division by the count, on
   // the derivative.
   CuVector<BaseFloat> counts(num_rows_out, kUndefined);
   if (num_log_count_features_ > 0) {
     counts.CopyColFromMat(out_value, 0);
     counts.ApplyExp();
   } else {
     counts.SetZero();
     // we need to recompute the counts from the input since they are not in the
     // output.  The submatrix initializer below takes num-rows, num-cols,
     // stride;  num-cols and stride are 1.
     CuSubMatrix<BaseFloat> counts_mat(counts.Data(), num_rows_out, 1, 1);
     counts_mat.AddRowRanges(in_value.ColRange(0, 1), indexes->forward_indexes);
   }
   // Divide the output derivative by the counts.  This is what we want as it
   // concerns the mean and x^2 stats.  As for the counts themselves, the
   // derivative will end up being discarded when we backprop to the
   // StatisticsExtractionComponent (as the count is not differentiable) so it
   // doesn't really matter.
   out_deriv.DivRowsVec(counts);
 
   // Now propagate the derivative back to the input.  we don't propagate it
   // back for the count's row since it's non-differentiable.
   in_deriv->ColRange(1, input_dim_ - 1).
       AddRowRanges(out_deriv.ColRange(num_log_count_features_, input_dim_ - 1),
                    indexes->backward_indexes);
 }

◆ Check()

void Check ( ) const

private

Definition at line 592 of file nnet-general-component.cc.

References StatisticsPoolingComponent::input_dim_, StatisticsPoolingComponent::input_period_, KALDI_ASSERT, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::output_stddevs_, StatisticsPoolingComponent::right_context_, and StatisticsPoolingComponent::variance_floor_.

Referenced by StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::StatisticsPoolingComponent().

                                              {
   KALDI_ASSERT(input_dim_ > 0);
   KALDI_ASSERT(input_period_ > 0);
   KALDI_ASSERT(left_context_ >= 0 && right_context_ >= 0 &&
                left_context_ + right_context_ > 0);
   KALDI_ASSERT(left_context_ % input_period_ == 0 &&
                right_context_ % input_period_ == 0);
   KALDI_ASSERT(variance_floor_ > 0.0 && variance_floor_ < 1.0);
   KALDI_ASSERT(!output_stddevs_ || (input_dim_ - 1) % 2 == 0);
 }

◆ Copy()

virtual Component* Copy ( ) const

inlinevirtual

Copies component (deep copy).

Implements Component.

Definition at line 374 of file nnet-general-component.h.

References DistributeComponent::GetInputIndexes(), DistributeComponent::IsComputable(), DistributeComponent::PrecomputeIndexes(), and Component::ReorderIndexes().

                                   {
     return new StatisticsPoolingComponent(*this);
   }

◆ GetInputIndexes()

void GetInputIndexes	(	const MiscComputationInfo &	misc_info,
		const Index &	output_index,
		std::vector< Index > *	desired_indexes
	)		const

virtual

This function only does something interesting for non-simple Components.

For a given index at the output of the component, tells us what indexes are required at its input (note: "required" encompasses also optionally-required things; it will enumerate all things that we'd like to have). See also IsComputable().

Parameters

[in]	misc_info	This argument is supplied to handle things that the framework can't very easily supply: information like which time indexes are needed for AggregateComponent, which time-indexes are available at the input of a recurrent network, and so on. We will add members to misc_info as needed.
[in]	output_index	The Index at the output of the component, for which we are requesting the list of indexes at the component's input.
[out]	desired_indexes	A list of indexes that are desired at the input. are to be written to here. By "desired" we mean required or optionally-required.

The default implementation of this function is suitable for any SimpleComponent; it just copies the output_index to a single identical element in input_indexes.

Reimplemented from Component.

Definition at line 654 of file nnet-general-component.cc.

References StatisticsPoolingComponent::input_period_, KALDI_ASSERT, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::right_context_, and Index::t.

                                              {
   desired_indexes->clear();
   Index input_index(output_index);
   int32 middle_t = output_index.t,
       t_start = middle_t - left_context_,
       t_last = middle_t + right_context_;
   KALDI_ASSERT(middle_t % input_period_ == 0);
   for (int32 t = t_start; t <= t_last; t += input_period_) {
     input_index.t = t;
     desired_indexes->push_back(input_index);
   }
 }

◆ 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 555 of file nnet-general-component.cc.

References StatisticsExtractionComponent::Check(), ConfigLine::GetValue(), ConfigLine::HasUnusedValues(), StatisticsExtractionComponent::input_dim_, StatisticsExtractionComponent::input_period_, KALDI_ERR, StatisticsExtractionComponent::Type(), ConfigLine::UnusedValues(), and ConfigLine::WholeLine().

                                                                {
   bool ok = cfl->GetValue("input-dim", &input_dim_);
   cfl->GetValue("input-period", &input_period_);
   cfl->GetValue("left-context", &left_context_);
   cfl->GetValue("right-context", &right_context_);
   cfl->GetValue("num-log-count-features", &num_log_count_features_);
   cfl->GetValue("output-stddevs", &output_stddevs_);
   cfl->GetValue("variance-floor", &variance_floor_);
 
   if (cfl->HasUnusedValues())
     KALDI_ERR << "Could not process these elements in initializer: "
               << cfl->UnusedValues();
   // do some basic checks here but Check() will check more completely.
   if (!ok || input_dim_ <= 0 || left_context_ + right_context_ <= 0 ||
       num_log_count_features_ < 0)
     KALDI_ERR << "Invalid initializer for layer of type "
               << Type() << ": \"" << cfl->WholeLine() << "\"";
   Check();
 }

◆ InputDim()

virtual int32 InputDim ( ) const

inlinevirtual

Returns input-dimension of this component.

Implements Component.

Definition at line 345 of file nnet-general-component.h.

References DistributeComponent::input_dim_.

345 { return input_dim_; }

kaldi::nnet3::StatisticsPoolingComponent::input_dim_

int32 input_dim_

Definition: nnet-general-component.h:413

◆ IsComputable()

bool IsComputable	(	const MiscComputationInfo &	misc_info,
		const Index &	output_index,
		const IndexSet &	input_index_set,
		std::vector< Index > *	used_inputs
	)		const

virtual

This function only does something interesting for non-simple Components, and it exists to make it possible to manage optionally-required inputs.

It tells the user whether a given output index is computable from a given set of input indexes, and if so, says which input indexes will be used in the computation.

Implementations of this function are required to have the property that adding an element to "input_index_set" can only ever change IsComputable from false to true, never vice versa.

Parameters

[in]	misc_info	Some information specific to the computation, such as minimum and maximum times for certain components to do adaptation on; it's a place to put things that don't easily fit in the framework.
[in]	output_index	The index that is to be computed at the output of this Component.
[in]	input_index_set	The set of indexes that is available at the input of this Component.
[out]	used_inputs	If this is non-NULL and the output is computable this will be set to the list of input indexes that will actually be used in the computation.

Returns: Returns true iff this output is computable from the provided inputs.

The default implementation of this function is suitable for any SimpleComponent: it just returns true if output_index is in input_index_set, and if so sets used_inputs to vector containing that one Index.

Reimplemented from Component.

Definition at line 670 of file nnet-general-component.cc.

References StatisticsPoolingComponent::input_period_, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::right_context_, and Index::t.

                                          {
   if (used_inputs)
     used_inputs->clear();
   // you are not supposed to access the output of this component other than at
   // multiples of the input period.  We could make this an error but decided to
   // just have it return false.
   if (output_index.t % input_period_ != 0)
     return false;
 
   Index input_index(output_index);
   int32 output_t = output_index.t,
       t_start = output_t - left_context_,
       t_last = output_t + right_context_;
   if (!used_inputs) {
     for (int32 t = t_start; t <= t_last; t += input_period_) {
       input_index.t = t;
       if (input_index_set(input_index))
         return true;
     }
     return false;
   } else {
     bool ans = false;
     for (int32 t = t_start; t <= t_last; t += input_period_) {
       input_index.t = t;
       if (input_index_set(input_index)) {
         ans = true;
         used_inputs->push_back(input_index);
       }
     }
     return ans;
   }
 }

◆ operator=()

StatisticsPoolingComponent& operator= ( const StatisticsPoolingComponent & other )

private

◆ OutputDim()

virtual int32 OutputDim ( ) const

inlinevirtual

Returns output-dimension of this component.

Implements Component.

Definition at line 346 of file nnet-general-component.h.

References DistributeComponent::InitFromConfig(), and DistributeComponent::input_dim_.

Referenced by ConstantComponent::Info(), and StatisticsPoolingComponent::Propagate().

                                   {
     return input_dim_ + num_log_count_features_ - 1;
   }

◆ PrecomputeIndexes()

ComponentPrecomputedIndexes * PrecomputeIndexes	(	const MiscComputationInfo &	misc_info,
		const std::vector< Index > &	input_indexes,
		const std::vector< Index > &	output_indexes,
		bool	need_backprop
	)		const

virtual

This function must return NULL for simple Components.

Returns a pointer to a class that may contain some precomputed component-specific and computation-specific indexes to be in used in the Propagate and Backprop functions.

Parameters

[in]	misc_info	This argument is supplied to handle things that the framework can't very easily supply: information like which time indexes are needed for AggregateComponent, which time-indexes are available at the input of a recurrent network, and so on. misc_info may not even ever be used here. We will add members to misc_info as needed.
[in]	input_indexes	A vector of indexes that explains what time-indexes (and other indexes) each row of the in/in_value/in_deriv matrices given to Propagate and Backprop will mean.
[in]	output_indexes	A vector of indexes that explains what time-indexes (and other indexes) each row of the out/out_value/out_deriv matrices given to Propagate and Backprop will mean.
[in]	need_backprop	True if we might need to do backprop with this component, so that if any different indexes are needed for backprop then those should be computed too.

Returns: Returns a child-class of class ComponentPrecomputedIndexes, or NULL if this component for does not need to precompute any indexes (e.g. if it is a simple component and does not care about indexes).

Reimplemented from Component.

Definition at line 708 of file nnet-general-component.cc.

References StatisticsPoolingComponentPrecomputedIndexes::backward_indexes, Int32Pair::first, StatisticsPoolingComponentPrecomputedIndexes::forward_indexes, rnnlm::i, StatisticsPoolingComponent::input_period_, KALDI_ASSERT, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::right_context_, Int32Pair::second, and Index::t.

                               {
   int32 num_input_indexes = input_indexes.size(),
       num_output_indexes = output_indexes.size();
   StatisticsPoolingComponentPrecomputedIndexes *ans = new
       StatisticsPoolingComponentPrecomputedIndexes();
 
   Int32Pair invalid_pair;
   invalid_pair.first = -1;
   invalid_pair.second = -1;
   // forward_indexes_cpu[i] will be the (begin, end) of input indexes
   // included in the sum for the i'th output index.
   std::vector<Int32Pair> forward_indexes_cpu(num_output_indexes,
                                              invalid_pair);
   // backward_indexes_cpu[i] will be the (begin, end) of output indexes
   // for which the i'th input index participates in the sum.
   // because of the way the indexes are sorted (and the fact that only
   // required indexes are present at the input), it naturally has this
   // structure [i.e. no gaps in the sets of indexes].
   std::vector<Int32Pair> backward_indexes_cpu(num_input_indexes,
                                               invalid_pair);
 
   // this map maps from Index to the position in 'input_indexes'.
   unordered_map<Index, int32, IndexHasher> index_to_input_pos;
   for (int32 i = 0; i < num_input_indexes; i++)
     index_to_input_pos[input_indexes[i]] = i;
 
   for (int32 i = 0; i < num_output_indexes; i++) {
     Index input_index(output_indexes[i]);
     int32 middle_t = input_index.t,
         t_start = middle_t - left_context_,
         t_last = middle_t + right_context_;
     for (int32 t = t_start; t <= t_last; t += input_period_) {
       input_index.t = t;
       unordered_map<Index, int32, IndexHasher>::iterator iter =
           index_to_input_pos.find(input_index);
       if (iter != index_to_input_pos.end()) {
         int32 input_pos = iter->second;
         if (forward_indexes_cpu[i].first == -1) {
           forward_indexes_cpu[i].first = input_pos;
           forward_indexes_cpu[i].second = input_pos + 1;
         } else {
           KALDI_ASSERT(forward_indexes_cpu[i].second == input_pos);
           forward_indexes_cpu[i].second++;
         }
         if (backward_indexes_cpu[input_pos].first == -1) {
           backward_indexes_cpu[input_pos].first = i;
           backward_indexes_cpu[input_pos].second = i + 1;
         } else {
           KALDI_ASSERT(backward_indexes_cpu[input_pos].second == i);
           backward_indexes_cpu[input_pos].second++;
         }
       }
     }
     KALDI_ASSERT(forward_indexes_cpu[i].first != -1);
   }
   for (int32 i = 0; i < num_input_indexes; i++) {
     KALDI_ASSERT(backward_indexes_cpu[i].first != -1);
   }
 
   ans->forward_indexes = forward_indexes_cpu;
   if (need_backprop)
     ans->backward_indexes = backward_indexes_cpu;
   return ans;
 }

◆ 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 777 of file nnet-general-component.cc.

                                         {
   out->SetZero();
   KALDI_ASSERT(indexes_in != NULL);
   const StatisticsPoolingComponentPrecomputedIndexes *indexes =
       dynamic_cast<const StatisticsPoolingComponentPrecomputedIndexes*>(indexes_in);
   int32 num_rows_out = out->NumRows();
   KALDI_ASSERT(indexes != NULL &&
                indexes->forward_indexes.Dim() == num_rows_out &&
                in.NumCols() == input_dim_ &&
                out->NumCols() == OutputDim());
   CuVector<BaseFloat> counts(num_rows_out);
   // counts_mat is a fake matrix with one column, containing the counts.
   CuSubMatrix<BaseFloat> counts_mat(counts.Data(), num_rows_out, 1, 1);
   counts_mat.AddRowRanges(in.ColRange(0, 1), indexes->forward_indexes);
 
   CuSubMatrix<BaseFloat> out_non_count(*out, 0, num_rows_out,
                                        num_log_count_features_, input_dim_ - 1);
   out_non_count.AddRowRanges(in.ColRange(1, input_dim_ - 1),
                              indexes->forward_indexes);
   out_non_count.DivRowsVec(counts);
 
   if (num_log_count_features_ > 0) {
     counts.ApplyLog();
     CuVector<BaseFloat> ones(num_log_count_features_, kUndefined);
     ones.Set(1.0);
     out->ColRange(0, num_log_count_features_).AddVecVec(1.0, counts, ones);
   }
 
   if (output_stddevs_) {
     // if this is true, then we assume the input contains x^2 stats as well as x
     // stats, and we want to process them into a standard deviation.
     KALDI_ASSERT((input_dim_ - 1) % 2 == 0);
     int32 feature_dim = (input_dim_ - 1) / 2;
     CuSubMatrix<BaseFloat> mean(*out, 0, num_rows_out,
                                 num_log_count_features_, feature_dim),
         variance(*out, 0, num_rows_out,
                  num_log_count_features_ + feature_dim, feature_dim);
     // subtract mean-squared from average of x^2 to get the variance.
     variance.AddMatMatElements(-1.0, mean, mean, 1.0);
     variance.ApplyFloor(variance_floor_);
     // compute the standard deviation via square root.
     variance.ApplyPow(0.5);
   }
   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 351 of file nnet-general-component.h.

References DistributeComponent::Backprop(), kaldi::nnet3::kBackpropAdds, kaldi::nnet3::kBackpropNeedsInput, kaldi::nnet3::kBackpropNeedsOutput, kaldi::nnet3::kReordersIndexes, DistributeComponent::Propagate(), DistributeComponent::Read(), and DistributeComponent::Write().

                                    {
     return kReordersIndexes|kBackpropAdds|
         (output_stddevs_ || num_log_count_features_ > 0 ?
          kBackpropNeedsOutput : 0) |
         (num_log_count_features_ == 0 ? kBackpropNeedsInput : 0);
   }

◆ 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 603 of file nnet-general-component.cc.

References StatisticsPoolingComponent::Check(), kaldi::ExpectOneOrTwoTokens(), kaldi::nnet3::ExpectToken(), StatisticsPoolingComponent::input_dim_, StatisticsPoolingComponent::input_period_, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::num_log_count_features_, StatisticsPoolingComponent::output_stddevs_, kaldi::ReadBasicType(), StatisticsPoolingComponent::require_direct_input_, StatisticsPoolingComponent::right_context_, and StatisticsPoolingComponent::variance_floor_.

                                                                  {
   ExpectOneOrTwoTokens(is, binary, "<StatisticsPoolingComponent>",
                        "<InputDim>");
   ReadBasicType(is, binary, &input_dim_);
   ExpectToken(is, binary, "<InputPeriod>");
   ReadBasicType(is, binary, &input_period_);
   ExpectToken(is, binary, "<LeftContext>");
   ReadBasicType(is, binary, &left_context_);
   ExpectToken(is, binary, "<RightContext>");
   ReadBasicType(is, binary, &right_context_);
   ExpectToken(is, binary, "<NumLogCountFeatures>");
   ReadBasicType(is, binary, &num_log_count_features_);
   ExpectToken(is, binary, "<OutputStddevs>");
   ReadBasicType(is, binary, &output_stddevs_);
   ExpectToken(is, binary, "<VarianceFloor>");
   ReadBasicType(is, binary, &variance_floor_);
   ExpectToken(is, binary, "</StatisticsPoolingComponent>");
   require_direct_input_ = false;  // This is not written to disk, it's only used
                                   // temporarily, in memory (see
                                   // nnet3-xvector-compute-batched.cc).
   Check();
 }

◆ ReorderIndexes()

void ReorderIndexes	(	std::vector< Index > *	input_indexes,
		std::vector< Index > *	output_indexes
	)		const

virtual

This function only does something interesting for non-simple Components.

It provides an opportunity for a Component to reorder the or pad the indexes at its input and output. This might be useful, for instance, if a component requires a particular ordering of the indexes that doesn't correspond to their natural ordering. Components that might modify the indexes are required to return the kReordersIndexes flag in their Properties(). The ReorderIndexes() function is now allowed to insert blanks into the indexes. The 'blanks' must be of the form (n,kNoTime,x), where the marker kNoTime (a very negative number) is there where the 't' indexes normally live. The reason we don't just have, say, (-1,-1,-1), relates to the need to preserve a regular pattern over the 'n' indexes so that 'shortcut compilation' (c.f. ExpandComputation()) can work correctly

Parameters

[in,out]	Indexes	at the input of the Component.
[in,out]	Indexes	at the output of the Component

Reimplemented from Component.

Definition at line 645 of file nnet-general-component.cc.

                                             {
     std::sort(input_indexes->begin(), input_indexes->end(),
               IndexLessNxt());
     std::sort(output_indexes->begin(), output_indexes->end(),
               IndexLessNxt());
 }

◆ SetRequireDirectInput()

void SetRequireDirectInput ( bool b )

inline

Definition at line 403 of file nnet-general-component.h.

403 { require_direct_input_ = b; }

kaldi::nnet3::StatisticsPoolingComponent::require_direct_input_

bool require_direct_input_

Definition: nnet-general-component.h:425

◆ Type()

virtual std::string Type ( ) const

inlinevirtual

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

Implements Component.

Definition at line 350 of file nnet-general-component.h.

Referenced by BackpropTruncationComponent::Info(), ConstantComponent::Info(), and BackpropTruncationComponent::InitFromConfig().

350 { return "StatisticsPoolingComponent"; }

◆ Write()

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

virtual

Write component to stream.

Implements Component.

Definition at line 626 of file nnet-general-component.cc.

References StatisticsPoolingComponent::input_dim_, StatisticsPoolingComponent::input_period_, StatisticsPoolingComponent::left_context_, StatisticsPoolingComponent::num_log_count_features_, StatisticsPoolingComponent::output_stddevs_, StatisticsPoolingComponent::right_context_, StatisticsPoolingComponent::variance_floor_, kaldi::WriteBasicType(), and kaldi::WriteToken().

                                                                         {
   WriteToken(os, binary, "<StatisticsPoolingComponent>");
   WriteToken(os, binary, "<InputDim>");
   WriteBasicType(os, binary, input_dim_);
   WriteToken(os, binary, "<InputPeriod>");
   WriteBasicType(os, binary, input_period_);
   WriteToken(os, binary, "<LeftContext>");
   WriteBasicType(os, binary, left_context_);
   WriteToken(os, binary, "<RightContext>");
   WriteBasicType(os, binary, right_context_);
   WriteToken(os, binary, "<NumLogCountFeatures>");
   WriteBasicType(os, binary, num_log_count_features_);
   WriteToken(os, binary, "<OutputStddevs>");
   WriteBasicType(os, binary, output_stddevs_);
   WriteToken(os, binary, "<VarianceFloor>");
   WriteBasicType(os, binary, variance_floor_);
   WriteToken(os, binary, "</StatisticsPoolingComponent>");
 }

Member Data Documentation

◆ input_dim_

int32 input_dim_

private

Definition at line 413 of file nnet-general-component.h.

Referenced by StatisticsPoolingComponent::Backprop(), StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::Propagate(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ input_period_

int32 input_period_

private

Definition at line 414 of file nnet-general-component.h.

Referenced by StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::GetInputIndexes(), StatisticsPoolingComponent::IsComputable(), StatisticsPoolingComponent::PrecomputeIndexes(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ left_context_

int32 left_context_

private

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

Referenced by StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::GetInputIndexes(), StatisticsPoolingComponent::IsComputable(), StatisticsPoolingComponent::PrecomputeIndexes(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ num_log_count_features_

int32 num_log_count_features_

private

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

Referenced by StatisticsPoolingComponent::Backprop(), StatisticsPoolingComponent::Propagate(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ output_stddevs_

bool output_stddevs_

private

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

Referenced by StatisticsPoolingComponent::Backprop(), StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::Propagate(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ require_direct_input_

bool require_direct_input_

private

Definition at line 425 of file nnet-general-component.h.

Referenced by StatisticsPoolingComponent::Read().

◆ right_context_

int32 right_context_

private

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

Referenced by StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::GetInputIndexes(), StatisticsPoolingComponent::IsComputable(), StatisticsPoolingComponent::PrecomputeIndexes(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

◆ variance_floor_

BaseFloat variance_floor_

private

Definition at line 419 of file nnet-general-component.h.

Referenced by StatisticsPoolingComponent::Check(), StatisticsPoolingComponent::Propagate(), StatisticsPoolingComponent::Read(), and StatisticsPoolingComponent::Write().

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

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

Public Member Functions

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ StatisticsPoolingComponent() [1/2]

◆ StatisticsPoolingComponent() [2/2]

Member Function Documentation

◆ Backprop()

◆ Check()

◆ Copy()

◆ GetInputIndexes()

◆ InitFromConfig()

◆ InputDim()

◆ IsComputable()

◆ operator=()

◆ OutputDim()

◆ PrecomputeIndexes()

◆ Propagate()

◆ Properties()

◆ Read()

◆ ReorderIndexes()

◆ SetRequireDirectInput()

◆ Type()

◆ Write()

Member Data Documentation

◆ input_dim_

◆ input_period_

◆ left_context_

◆ num_log_count_features_

◆ output_stddevs_

◆ require_direct_input_

◆ right_context_

◆ variance_floor_