C++ API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2

List of all members
Distributed< step2Master, algorithmFPType, method > Class Template Reference

Provides methods for neural network model-based training in the batch processing mode. More...

Class Declaration

template<typename algorithmFPType, Method method>
class daal::algorithms::neural_networks::training::interface1::Distributed< step2Master, algorithmFPType, method >

Template Parameters
algorithmFPTypeData type to use in intermediate computations for neural network model-based training, double or float
methodNeural network training method, training::Method
Enumerations
References

Constructor & Destructor Documentation

◆ Distributed()

Distributed ( const Distributed< step2Master, algorithmFPType, method > &  other)
inline

Constructs neural network by copying input objects and parameters of another neural network

Parameters
[in]otherAn algorithm to be used as the source to initialize the input objects and parameters of the algorithm

Member Function Documentation

◆ clone()

services::SharedPtr<Distributed<step2Master, algorithmFPType, method> > clone ( ) const
inline

Returns a pointer to the newly allocated neural network with a copy of input objects and parameters of this neural network

Returns
Pointer to the newly allocated layer

◆ getMethod()

virtual int getMethod ( ) const
inlinevirtual

Returns method of the algorithm

Returns
Method of the algorithm

Implements AlgorithmIface.

◆ getPartialResult()

DistributedPartialResultPtr getPartialResult ( )
inline

Returns structure that contains computed partial results of the neural network algorithm

Returns
Structure that contains partial results of the neural network algorithm

◆ getResult()

ResultPtr getResult ( )
inline

Returns the structure that contains the results of the neural network algorithm

Returns
Structure that contains the results of the neural network algorithm

◆ initialize()

services::Status initialize ( const services::Collection< size_t > &  dataSize,
const training::Topology topology 
)
inline

Initializes the neural network topology

Parameters
[in]dataSizeCollection of sizes of each dimension of input data tensor
[in]topologyNeural network topology
Returns
Status of computations

◆ setPartialResult()

services::Status setPartialResult ( const DistributedPartialResultPtr partialResult)
inline

Registers user-allocated memory to store partial results of the neural network algorithm

Parameters
[in]partialResultStructure for storing partial results of the neural network algorithm
Returns
Status of computations

Member Data Documentation

◆ input

Input data structure

◆ parameter

ParameterType parameter

Training parameters


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

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