Python* API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1

Public Member Functions | Static Public Attributes | List of all members
Distributed Class Reference

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

Public Member Functions

def __init__
 
def setPartialResult
 
def getPartialResult
 
def getResult
 
def setResult
 
def clone
 
def getMethod
 
def compute
 
def finalizeCompute
 

Static Public Attributes

 input = ...
 
 parameter = ...
 

Detailed Description

Parameters
fptypeData type to use in intermediate computations for neural network model-based training, double or float
methodNeural network training method, training.Method
Enumerations
  • training.Method Computation methods
References
Deprecated:
This item will be removed in a future release.
Aliases
  • Distributed_Step1LocalFloat64FeedforwardDense is an alias of Distributed(step=daal.step1Local, fptype=float64, method=daal.algorithms.neural_networks.training.feedforwardDense)
  • Distributed_Step1LocalFloat32FeedforwardDense is an alias of Distributed(step=daal.step1Local, fptype=float32, method=daal.algorithms.neural_networks.training.feedforwardDense)
  • Distributed_Step2MasterFloat64FeedforwardDense is an alias of Distributed(step=daal.step2Master, fptype=float64, method=daal.algorithms.neural_networks.training.feedforwardDense)
  • Distributed_Step2MasterFloat32FeedforwardDense is an alias of Distributed(step=daal.step2Master, fptype=float32, method=daal.algorithms.neural_networks.training.feedforwardDense)

Constructor & Destructor Documentation

def __init__ (   self,
  args 
)

Variant 1
Default constructor
Deprecated:
This item will be removed in a future release.

Variant 2

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

Parameters
otherAn algorithm to be used as the source to initialize the input objects and parameters of the algorithm
Deprecated:
This item will be removed in a future release.

Member Function Documentation

def clone (   self)

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
Deprecated:
This item will be removed in a future release.
def compute (   self)

Invokes computations and returns partial result

def finalizeCompute (   self)

Finalizes computations and returns (final) result

def getMethod (   self)

Returns method of the algorithm

Returns
Method of the algorithm
Deprecated:
This item will be removed in a future release.
def getPartialResult (   self)

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

Returns
Structure that contains partial results of the neural network algorithm
Deprecated:
This item will be removed in a future release.
def getResult (   self)

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

Returns
Structure that contains the results of the neural network algorithm
Deprecated:
This item will be removed in a future release.
def setPartialResult (   self,
  partialResult 
)

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

Parameters
partialResultStructure for storing partial results of the neural network algorithm
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
def setResult (   self,
  res 
)

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

Parameters
resStructure for storing results of the neural network algorithm
Returns
Status of computations
Deprecated:
This item will be removed in a future release.

Member Data Documentation

input = ...
static

Input data structure

parameter = ...
static

Training parameters


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

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