Java* API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2

List of all members
TrainingBatch Class Reference

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

Class Constructor

◆ TrainingBatch() [1/4]

TrainingBatch ( DaalContext  context,
Class<? extends Number >  cls,
TrainingMethod  method,
Batch  optimizationSolver 
)

Constructs neural network model-based training object

Parameters
contextContext to manage the neural network training object
clsData type to use in intermediate computations for the neural network, Double.class or Float.class
methodNeural network computation method, TrainingMethod
optimizationSolverOptimization solver

◆ TrainingBatch() [2/4]

TrainingBatch ( DaalContext  context,
Class<? extends Number >  cls,
Batch  optimizationSolver 
)

Constructs neural network model-based training object with default computation method

Parameters
contextContext to manage the neural network training object
clsData type to use in intermediate computations for the neural network, Double.class or Float.class
optimizationSolverOptimization solver

◆ TrainingBatch() [3/4]

TrainingBatch ( DaalContext  context,
Batch  optimizationSolver 
)

Constructs neural network model-based training object with Float data type used for intermediate computations and default computation method

Parameters
contextContext to manage the neural network training object
optimizationSolverOptimization solver

◆ TrainingBatch() [4/4]

TrainingBatch ( DaalContext  context,
TrainingBatch  other 
)

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

Parameters
contextContext to manage the neural network
otherA neural network to be used as the source to initialize the input objects and parameters of the neural network

Detailed Description

Member Function Documentation

◆ clone()

TrainingBatch clone ( DaalContext  context)

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

Parameters
contextContext to manage the neural network
Returns
The newly allocated neural network

◆ compute()

TrainingResult compute ( )

Runs the neural network in the batch processing mode

Returns
Results of the neural network in the batch processing mode

◆ initialize()

void initialize ( long []  dataSize,
TrainingTopology  topology 
)

Initializes the neural network topology of the layers

Parameters
dataSizeSize of the input data for the trainings
topologyTrainingTopology of the layers

◆ setResult()

void setResult ( TrainingResult  result)

Registers user-allocated memory to store the results of the neural network in the batch processing mode

Parameters
resultStructure for storing the results of the neural network

Member Data Documentation

◆ input

Input data structure

◆ method

Neural network training method

◆ parameter

TrainingParameter parameter

Training parameters

◆ prec

Precision prec
protected

Data type to use in intermediate computations for neural network model-based training


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

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