RepeatedAffineComponent Class Reference

#include <nnet-simple-component.h>

Inheritance diagram for RepeatedAffineComponent:
Collaboration diagram for RepeatedAffineComponent:

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...
 
 RepeatedAffineComponent ()
 
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 ComponentCopy () 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...
 
const CuVector< BaseFloat > & BiasParams () const
 
const CuMatrix< BaseFloat > & LinearParams () const
 
 RepeatedAffineComponent (const RepeatedAffineComponent &other)
 
friend BlockAffineComponent::BlockAffineComponent (const RepeatedAffineComponent &rac)
 
- 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 ComponentPrecomputedIndexesPrecomputeIndexes (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 (int32 input_dim, int32 output_dim, int32 num_repeats, BaseFloat param_stddev, BaseFloat bias_mean, BaseFloat bias_stddev)
 
virtual void Update (const CuMatrixBase< BaseFloat > &in_value, const CuMatrixBase< BaseFloat > &out_deriv)
 
virtual void SetNaturalGradientConfigs ()
 
const RepeatedAffineComponentoperator= (const RepeatedAffineComponent &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< BaseFloatlinear_params_
 
CuVector< BaseFloatbias_params_
 
int32 num_repeats_
 
- 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...
 

Additional Inherited Members

- Static Public Member Functions inherited from Component
static ComponentReadNew (std::istream &is, bool binary)
 Read component from stream (works out its type). Dies on error. More...
 
static ComponentNewComponentOfType (const std::string &type)
 Returns a new Component of the given type e.g. More...
 

Detailed Description

Definition at line 570 of file nnet-simple-component.h.

Constructor & Destructor Documentation

◆ RepeatedAffineComponent() [1/2]

Definition at line 579 of file nnet-simple-component.h.

Referenced by RepeatedAffineComponent::Copy().

579 { } // use Init to really initialize.

◆ RepeatedAffineComponent() [2/2]

RepeatedAffineComponent ( const RepeatedAffineComponent other)
explicit

Definition at line 1339 of file nnet-simple-component.cc.

1339  :
1340  UpdatableComponent(component),
1341  linear_params_(component.linear_params_),
1342  bias_params_(component.bias_params_),
1343  num_repeats_(component.num_repeats_) {}

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 1356 of file nnet-simple-component.cc.

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

1356  {
1357  const RepeatedAffineComponent *other =
1358  dynamic_cast<const RepeatedAffineComponent *>(&other_in);
1359  KALDI_ASSERT(other != NULL);
1360  linear_params_.AddMat(alpha, other->linear_params_);
1361  bias_params_.AddVec(alpha, other->bias_params_);
1362 }
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185

◆ 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_infoThe component name, to be printed out in any warning messages.
[in]indexesA 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_valueThe matrix that was given as input to the Propagate function. Will be ignored (and may be empty) if Properties()&kBackpropNeedsInput == 0.
[in]out_valueThe matrix that was output from the Propagate function. Will be ignored (and may be empty) if Properties()&kBackpropNeedsOutput == 0
[in]out_derivThe derivative at the output of this component.
[in]memoThis 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_updateIf 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_derivThe 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 1463 of file nnet-simple-component.cc.

References CuMatrixBase< Real >::AddMatMat(), CuMatrixBase< Real >::Data(), KALDI_ASSERT, kaldi::kNoTrans, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, CuMatrixBase< Real >::NumCols(), CuMatrixBase< Real >::NumRows(), NVTX_RANGE, CuMatrixBase< Real >::Stride(), and RepeatedAffineComponent::Update().

1470  {
1471  NVTX_RANGE("RepeatedAffineComponent::Backprop");
1472  KALDI_ASSERT(out_deriv.NumCols() == out_deriv.Stride() &&
1473  (in_value.NumCols() == 0 || in_value.NumCols() == in_value.Stride()) &&
1474  (!in_deriv || in_deriv->NumCols() == in_deriv->Stride()));
1475 
1476  RepeatedAffineComponent *to_update = dynamic_cast<RepeatedAffineComponent*>(
1477  to_update_in);
1478 
1479  // Propagate the derivative back to the input.
1480  // add with coefficient 1.0 since property kBackpropAdds is true.
1481  // If we wanted to add with coefficient 0.0 we'd need to zero the
1482  // in_deriv, in case of infinities.
1483  if (in_deriv) {
1484  int32 num_repeats = num_repeats_,
1485  num_rows = out_deriv.NumRows(),
1486  block_dim_out = linear_params_.NumRows(),
1487  block_dim_in = linear_params_.NumCols();
1488 
1489  CuSubMatrix<BaseFloat> in_deriv_reshaped(in_deriv->Data(),
1490  num_rows * num_repeats,
1491  block_dim_in, block_dim_in),
1492  out_deriv_reshaped(out_deriv.Data(),
1493  num_rows * num_repeats,
1494  block_dim_out, block_dim_out);
1495  in_deriv_reshaped.AddMatMat(1.0, out_deriv_reshaped, kNoTrans,
1496  linear_params_, kNoTrans, 1.0);
1497  }
1498 
1499  // Next update the model (must do this 2nd so the derivatives we propagate are
1500  // accurate, in case this == to_update_in.)
1501  if (to_update != NULL)
1502  to_update->Update(in_value, out_deriv);
1503 }
kaldi::int32 int32
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185
#define NVTX_RANGE(name)
Definition: cu-common.h:143

◆ BiasParams()

const CuVector<BaseFloat>& BiasParams ( ) const
inline

Definition at line 612 of file nnet-simple-component.h.

612 { return bias_params_; }

◆ BlockAffineComponent::BlockAffineComponent()

◆ Copy()

Component * Copy ( ) const
virtual

Copies component (deep copy).

Implements Component.

Reimplemented in NaturalGradientRepeatedAffineComponent.

Definition at line 1382 of file nnet-simple-component.cc.

References RepeatedAffineComponent::RepeatedAffineComponent().

1382  {
1384  return ans;
1385 }

◆ 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 1387 of file nnet-simple-component.cc.

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

1387  {
1388  const RepeatedAffineComponent *other =
1389  dynamic_cast<const RepeatedAffineComponent*>(&other_in);
1390  return TraceMatMat(linear_params_, other->linear_params_, kTrans)
1391  + VecVec(bias_params_, other->bias_params_);
1392 }
Real TraceMatMat(const MatrixBase< Real > &A, const MatrixBase< Real > &B, MatrixTransposeType trans)
We need to declare this here as it will be a friend function.
Real VecVec(const VectorBase< Real > &a, const VectorBase< Real > &b)
Returns dot product between v1 and v2.
Definition: kaldi-vector.cc:37

◆ 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 1373 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, UpdatableComponent::Info(), RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, and kaldi::nnet3::PrintParameterStats().

1373  {
1374  std::ostringstream stream;
1375  stream << UpdatableComponent::Info()
1376  << ", num-repeats=" << num_repeats_;
1377  PrintParameterStats(stream, "linear-params", linear_params_);
1378  PrintParameterStats(stream, "bias", bias_params_, true);
1379  return stream.str();
1380 }
virtual std::string Info() const
Returns some text-form information about this component, for diagnostics.
void PrintParameterStats(std::ostringstream &os, const std::string &name, const CuVectorBase< BaseFloat > &params, bool include_mean)
Print to &#39;os&#39; some information about the mean and standard deviation of some parameters, used in Info() functions in nnet-simple-component.cc.
Definition: nnet-parse.cc:157

◆ Init()

void Init ( int32  input_dim,
int32  output_dim,
int32  num_repeats,
BaseFloat  param_stddev,
BaseFloat  bias_mean,
BaseFloat  bias_stddev 
)
protected

Definition at line 1394 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, KALDI_ASSERT, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, and RepeatedAffineComponent::SetNaturalGradientConfigs().

Referenced by RepeatedAffineComponent::InitFromConfig().

1396  {
1397  KALDI_ASSERT(input_dim % num_repeats == 0 && output_dim % num_repeats == 0);
1398  linear_params_.Resize(output_dim / num_repeats, input_dim / num_repeats);
1399  bias_params_.Resize(output_dim / num_repeats);
1400  num_repeats_ = num_repeats;
1401  KALDI_ASSERT(output_dim > 0 && input_dim > 0 && param_stddev >= 0.0);
1402  linear_params_.SetRandn(); // sets to random normally distributed noise.
1403  linear_params_.Scale(param_stddev);
1404  bias_params_.SetRandn();
1405  bias_params_.Scale(bias_stddev);
1406  bias_params_.Add(bias_mean);
1408 }
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185

◆ InitFromConfig()

void InitFromConfig ( ConfigLine cfl)
virtual

Initialize, from a ConfigLine object.

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

Implements Component.

Definition at line 1411 of file nnet-simple-component.cc.

References ConfigLine::GetValue(), ConfigLine::HasUnusedValues(), RepeatedAffineComponent::Init(), UpdatableComponent::InitLearningRatesFromConfig(), KALDI_ASSERT, KALDI_ERR, RepeatedAffineComponent::num_repeats_, ConfigLine::UnusedValues(), and ConfigLine::WholeLine().

1411  {
1412  bool ok = true;
1413  int32 num_repeats = num_repeats_;
1414  int32 input_dim = -1, output_dim = -1;
1416  ok = cfl->GetValue("num-repeats", &num_repeats) && ok;
1417  ok = cfl->GetValue("input-dim", &input_dim) && ok;
1418  ok = cfl->GetValue("output-dim", &output_dim) && ok;
1419  KALDI_ASSERT(input_dim % num_repeats == 0 &&
1420  "num-repeats must divide input-dim");
1421  KALDI_ASSERT(output_dim % num_repeats == 0 &&
1422  "num-repeats must divide output-dim");
1423  BaseFloat param_stddev = 1.0 / std::sqrt(input_dim / num_repeats),
1424  bias_mean = 0.0, bias_stddev = 0.0;
1425  cfl->GetValue("param-stddev", &param_stddev);
1426  cfl->GetValue("bias-mean", &bias_mean);
1427  cfl->GetValue("bias-stddev", &bias_stddev);
1428  Init(input_dim, output_dim,
1429  num_repeats, param_stddev, bias_mean, bias_stddev);
1430  if (cfl->HasUnusedValues())
1431  KALDI_ERR << "Could not process these elements in initializer: "
1432  << cfl->UnusedValues();
1433  if (!ok)
1434  KALDI_ERR << "Bad initializer " << cfl->WholeLine();
1435 }
void InitLearningRatesFromConfig(ConfigLine *cfl)
kaldi::int32 int32
float BaseFloat
Definition: kaldi-types.h:29
#define KALDI_ERR
Definition: kaldi-error.h:147
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185
void Init(int32 input_dim, int32 output_dim, int32 num_repeats, BaseFloat param_stddev, BaseFloat bias_mean, BaseFloat bias_stddev)

◆ InputDim()

virtual int32 InputDim ( ) const
inlinevirtual

Returns input-dimension of this component.

Implements Component.

Definition at line 573 of file nnet-simple-component.h.

◆ LinearParams()

const CuMatrix<BaseFloat>& LinearParams ( ) const
inline

◆ 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 1563 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, and RepeatedAffineComponent::linear_params_.

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

1563  {
1564  // Note: unlike AffineComponent, InputDim() & OutputDim() are not used here and below,
1565  // for they are multipled by num_repeats_.
1566  return linear_params_.NumCols() * linear_params_.NumRows() + bias_params_.Dim();
1567 }

◆ operator=()

const RepeatedAffineComponent& operator= ( const RepeatedAffineComponent other)
protected

◆ OutputDim()

virtual int32 OutputDim ( ) const
inlinevirtual

Returns output-dimension of this component.

Implements Component.

Definition at line 574 of file nnet-simple-component.h.

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

◆ 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 1364 of file nnet-simple-component.cc.

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

1364  {
1365  CuMatrix<BaseFloat> temp_linear_params(linear_params_);
1366  temp_linear_params.SetRandn();
1367  linear_params_.AddMat(stddev, temp_linear_params);
1368  CuVector<BaseFloat> temp_bias_params(bias_params_);
1369  temp_bias_params.SetRandn();
1370  bias_params_.AddVec(stddev, temp_bias_params);
1371 }

◆ Propagate()

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

Propagate function.

Parameters
[in]indexesA 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]inThe input to this component. Num-columns == InputDim().
[out]outThe 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 1437 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, CuMatrixBase< Real >::CopyRowsFromVec(), CuMatrixBase< Real >::Data(), KALDI_ASSERT, kaldi::kNoTrans, kaldi::kTrans, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, CuMatrixBase< Real >::NumCols(), CuMatrixBase< Real >::NumRows(), and CuMatrixBase< Real >::Stride().

1439  {
1440  // we gave the kInputContiguous and kOutputContiguous flags-- check that they
1441  // are honored.
1442  KALDI_ASSERT(in.NumCols() == in.Stride() &&
1443  out->NumCols() == out->Stride() &&
1444  out->NumRows() == in.NumRows());
1445 
1446  int32 num_repeats = num_repeats_,
1447  num_rows = in.NumRows(),
1448  block_dim_out = linear_params_.NumRows(),
1449  block_dim_in = linear_params_.NumCols();
1450 
1451  CuSubMatrix<BaseFloat> in_reshaped(in.Data(), num_rows * num_repeats,
1452  block_dim_in, block_dim_in),
1453  out_reshaped(out->Data(), num_rows * num_repeats,
1454  block_dim_out, block_dim_out);
1455 
1456  out_reshaped.CopyRowsFromVec(bias_params_);
1457 
1458  out_reshaped.AddMatMat(1.0, in_reshaped, kNoTrans,
1459  linear_params_, kTrans, 1.0);
1460  return NULL;
1461 }
kaldi::int32 int32
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185

◆ Properties()

◆ 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 1531 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, kaldi::nnet3::ExpectToken(), UpdatableComponent::is_gradient_, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, kaldi::PeekToken(), kaldi::ReadBasicType(), UpdatableComponent::ReadUpdatableCommon(), RepeatedAffineComponent::SetNaturalGradientConfigs(), and RepeatedAffineComponent::Type().

1531  {
1532  // This Read function also works for NaturalGradientRepeatedAffineComponent.
1533  ReadUpdatableCommon(is, binary); // read opening tag and learning rate.
1534  ExpectToken(is, binary, "<NumRepeats>");
1535  ReadBasicType(is, binary, &num_repeats_);
1536  ExpectToken(is, binary, "<LinearParams>");
1537  linear_params_.Read(is, binary);
1538  ExpectToken(is, binary, "<BiasParams>");
1539  bias_params_.Read(is, binary);
1540  if (PeekToken(is, binary) == 'I') {
1541  // for back compatibility; we don't write this here any
1542  // more as it's written and read in Write/ReadUpdatableCommon
1543  ExpectToken(is, binary, "<IsGradient>");
1544  ReadBasicType(is, binary, &is_gradient_);
1545  }
1546  ExpectToken(is, binary, std::string("</") + Type() + std::string(">"));
1548 }
void ReadBasicType(std::istream &is, bool binary, T *t)
ReadBasicType is the name of the read function for bool, integer types, and floating-point types...
Definition: io-funcs-inl.h:55
static void ExpectToken(const std::string &token, const std::string &what_we_are_parsing, const std::string **next_token)
std::string ReadUpdatableCommon(std::istream &is, bool binary)
virtual std::string Type() const
Returns a string such as "SigmoidComponent", describing the type of the object.
int PeekToken(std::istream &is, bool binary)
PeekToken will return the first character of the next token, or -1 if end of file.
Definition: io-funcs.cc:170
bool is_gradient_
True if this component is to be treated as a gradient rather than as parameters.

◆ 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 1346 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, and RepeatedAffineComponent::linear_params_.

1346  {
1347  if (scale == 0.0) {
1348  linear_params_.SetZero();
1349  bias_params_.SetZero();
1350  } else {
1351  linear_params_.Scale(scale);
1352  bias_params_.Scale(scale);
1353  }
1354 }

◆ SetNaturalGradientConfigs()

virtual void SetNaturalGradientConfigs ( )
inlineprotectedvirtual

◆ Type()

virtual std::string Type ( ) const
inlinevirtual

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

Implements Component.

Reimplemented in NaturalGradientRepeatedAffineComponent.

Definition at line 580 of file nnet-simple-component.h.

Referenced by RepeatedAffineComponent::Read(), and RepeatedAffineComponent::Write().

580 { return "RepeatedAffineComponent"; }

◆ UnVectorize()

void UnVectorize ( const VectorBase< BaseFloat > &  params)
virtual

Converts the parameters from vector form.

Reimplemented from UpdatableComponent.

Definition at line 1576 of file nnet-simple-component.cc.

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

1576  {
1577  KALDI_ASSERT(params.Dim() == this->NumParameters());
1578  linear_params_.CopyRowsFromVec(params.Range(0, linear_params_.NumCols() * linear_params_.NumRows()));
1579  bias_params_.CopyFromVec(params.Range(linear_params_.NumCols() * linear_params_.NumRows(),
1580  bias_params_.Dim()));
1581 }
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185
virtual int32 NumParameters() const
The following new virtual function returns the total dimension of the parameters in this class...

◆ Update()

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

Reimplemented in NaturalGradientRepeatedAffineComponent.

Definition at line 1505 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, CuMatrixBase< Real >::Data(), KALDI_ASSERT, kaldi::kNoTrans, kaldi::kTrans, UpdatableComponent::learning_rate_, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, CuMatrixBase< Real >::NumCols(), CuMatrixBase< Real >::NumRows(), and CuMatrixBase< Real >::Stride().

Referenced by RepeatedAffineComponent::Backprop().

1506  {
1507  KALDI_ASSERT(out_deriv.NumCols() == out_deriv.Stride() &&
1508  in_value.NumCols() == in_value.Stride() &&
1509  in_value.NumRows() == out_deriv.NumRows());
1510 
1511 
1512  int32 num_repeats = num_repeats_,
1513  num_rows = in_value.NumRows(),
1514  block_dim_out = linear_params_.NumRows(),
1515  block_dim_in = linear_params_.NumCols();
1516 
1517  CuSubMatrix<BaseFloat> in_value_reshaped(in_value.Data(),
1518  num_rows * num_repeats,
1519  block_dim_in, block_dim_in),
1520  out_deriv_reshaped(out_deriv.Data(),
1521  num_rows * num_repeats,
1522  block_dim_out, block_dim_out);
1523 
1524 
1525  linear_params_.AddMatMat(learning_rate_, out_deriv_reshaped, kTrans,
1526  in_value_reshaped, kNoTrans, 1.0);
1527  bias_params_.AddRowSumMat(learning_rate_,
1528  out_deriv_reshaped);
1529 }
kaldi::int32 int32
BaseFloat learning_rate_
learning rate (typically 0.0..0.01)
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185

◆ 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 1569 of file nnet-simple-component.cc.

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

1569  {
1570  KALDI_ASSERT(params->Dim() == this->NumParameters());
1571  params->Range(0, linear_params_.NumCols() * linear_params_.NumRows()).CopyRowsFromMat(linear_params_);
1572  params->Range(linear_params_.NumCols() * linear_params_.NumRows(),
1573  bias_params_.Dim()).CopyFromVec(bias_params_);
1574 }
#define KALDI_ASSERT(cond)
Definition: kaldi-error.h:185
virtual int32 NumParameters() const
The following new virtual function returns the total dimension of the parameters in this class...

◆ Write()

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

Write component to stream.

Implements Component.

Definition at line 1550 of file nnet-simple-component.cc.

References RepeatedAffineComponent::bias_params_, RepeatedAffineComponent::linear_params_, RepeatedAffineComponent::num_repeats_, RepeatedAffineComponent::Type(), kaldi::WriteBasicType(), kaldi::WriteToken(), and UpdatableComponent::WriteUpdatableCommon().

1550  {
1551  // This Write function also works for NaturalGradientRepeatedAffineComponent.
1552  WriteUpdatableCommon(os, binary); // Write opening tag and learning rate
1553  WriteToken(os, binary, "<NumRepeats>");
1554  WriteBasicType(os, binary, num_repeats_);
1555  WriteToken(os, binary, "<LinearParams>");
1556  linear_params_.Write(os, binary);
1557  WriteToken(os, binary, "<BiasParams>");
1558  bias_params_.Write(os, binary);
1559  // write closing token.
1560  WriteToken(os, binary, std::string("</") + Type() + std::string(">"));
1561 }
virtual std::string Type() const
Returns a string such as "SigmoidComponent", describing the type of the object.
void WriteToken(std::ostream &os, bool binary, const char *token)
The WriteToken functions are for writing nonempty sequences of non-space characters.
Definition: io-funcs.cc:134
void WriteUpdatableCommon(std::ostream &is, bool binary) const
void WriteBasicType(std::ostream &os, bool binary, T t)
WriteBasicType is the name of the write function for bool, integer types, and floating-point types...
Definition: io-funcs-inl.h:34

Member Data Documentation

◆ bias_params_

◆ linear_params_

◆ num_repeats_


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