C++ API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1

List of all members
Parameter< momentum > Struct Template Reference

Parameter for the Stochastic gradient descent algorithm More...

Class Declaration

template<>
struct daal::algorithms::optimization_solver::sgd::interface2::Parameter< momentum >

template<>
struct DAAL_EXPORT Parameter<momentum> : public BaseParameter
{
const sum_of_functions::BatchPtr& function,
double momentum = 0.9,
size_t nIterations = 100,
double accuracyThreshold = 1.0e-05,
data_management::NumericTablePtr batchIndices = data_management::NumericTablePtr(),
size_t batchSize = 128,
data_management::NumericTablePtr learningRateSequence = data_management::NumericTablePtr(
new data_management::HomogenNumericTable<double>(
1, 1, data_management::NumericTableIface::doAllocate, 1.0)),
size_t seed = 777 );
virtual services::Status check() const;
virtual ~Parameter() {}
double momentum;
};

Constructor & Destructor Documentation

Parameter ( const sum_of_functions::BatchPtr &  function,
double  momentum = 0.9,
size_t  nIterations = 100,
double  accuracyThreshold = 1.0e-05,
data_management::NumericTablePtr  batchIndices = data_management::NumericTablePtr(),
size_t  batchSize = 128,
data_management::NumericTablePtr  learningRateSequence = data_management::NumericTablePtr(new data_management::HomogenNumericTable< double >(1, 1, data_management::NumericTableIface::doAllocate, 1.0)),
size_t  seed = 777 
)

Constructs the parameter class of the Stochastic gradient descent algorithm

Parameters
[in]functionObjective function represented as sum of functions
[in]momentumThe momentum value
[in]nIterationsMaximal number of iterations of the algorithm
[in]accuracyThresholdAccuracy of the algorithm. The algorithm terminates when this accuracy is achieved
[in]batchIndicesNumeric table that represents 32 bit integer indices of terms in the objective function. If no indices are provided, the implementation will generate random indices.
[in]batchSizeNumber of batch indices to compute the stochastic gradient. If batchSize is equal to the number of terms in objective function then no random sampling is performed, and all terms are used to calculate the gradient. This parameter is ignored if batchIndices is provided.
[in]learningRateSequenceNumeric table that contains values of the learning rate sequence
[in]seedSeed for random generation of 32 bit integer indices of terms in the objective function.
Deprecated:
This item will be removed in a future release. Use engine instead.

Member Function Documentation

virtual services::Status check ( ) const
virtual

Checks the correctness of the parameter

Returns
Status of computations

Reimplemented from BaseParameter.

Member Data Documentation

double momentum

Momentum value


The documentation for this struct was generated from the following file:

For more complete information about compiler optimizations, see our Optimization Notice.