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

Public Member Functions | List of all members
Model Class Reference

Class representing the model of neural network. More...

Public Member Functions

def serializationTag
 
def getSerializationTag
 
def create
 
def __init__
 Constructor.
 
def getForwardLayers
 
def getForwardLayer
 
def getBackwardLayers
 
def getBackwardLayer
 
def getWeightsAndBiasesStorageStatus
 
def setWeightsAndBiases
 
def getWeightsAndBiases
 
def getWeightsAndBiasesDerivatives
 
def setErrors
 
def getSolverOptionalArgument
 
def setSolverOptionalArgument
 
def getSolverOptionalArgumentCollection
 
def setSolverOptionalArgumentCollection
 
def allocate_{Float64|Float32}
 
def getPredictionModel_{Float64|Float32}
 
def initialize_{Float64|Float32}
 
- Public Member Functions inherited from ModelImpl
def getNextLayers
 
def setWeightsAndBiases
 
def getWeightsAndBiases
 
- Public Member Functions inherited from Model
def __init__
 
def getSerializationTag
 
- Public Member Functions inherited from SerializationIface
def serialize
 
def deserialize
 
def getSerializationTag
 
- Public Member Functions inherited from Base
def __init__
 

Detailed Description

Deprecated:
This item will be removed in a future release.

Constructor & Destructor Documentation

def __init__ (   self,
  args 
)

Variant 1

Variant 2
Copy constructor

Member Function Documentation

def allocate_{Float64|Float32} (   self,
  args 
)

Variant 1

Allocates the buffers needed for the training using neural network

Parameters
sampleSizeDimensionality of the batch for the input to the first layer
parameterParameters of the training
Returns
Status of computations
Deprecated:
This item will be removed in a future release.

Variant 2

Allocates the buffers needed for the training using neural network

Parameters
sampleSizeDimensionality of the batch for the input to the first layer
parameterParameters of the training
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
Full Names
  • allocate_Float64 is for float64
  • allocate_Float32 is for float32
def create (   stat = None)

create(Status stat=None) -> daal.services.SharedPtr< daal.algorithms.neural_networks.training.Model >

def getBackwardLayer (   self,
  index 
)

Returns the backward stage of a layer with certain index in the network

Parameters
indexIndex of the layer in the network
Returns
Backward stage of a layer with certain index in the network
Deprecated:
This item will be removed in a future release.
def getBackwardLayers (   self)

Returns list of backward layers

Returns
List of backward layers
Deprecated:
This item will be removed in a future release.
def getForwardLayer (   self,
  index 
)

Returns the forward stage of a layer with certain index in the network

Parameters
indexIndex of the layer in the network
Returns
Forward stage of a layer with certain index in the network
Deprecated:
This item will be removed in a future release.
def getForwardLayers (   self)

Returns list of forward layers

Returns
List of forward layers
Deprecated:
This item will be removed in a future release.
def getPredictionModel_{Float64|Float32} (   self)

Returns list of forward layers and their parameters organised in the prediction.Model

Returns
List of forward layers and their parameters organised in the prediction.Model
Deprecated:
This item will be removed in a future release.
Full Names
  • getPredictionModel_Float64 is for float64
  • getPredictionModel_Float32 is for float32
def getSerializationTag (   self)

getSerializationTag(Model self) -> int

def getSolverOptionalArgument (   self,
  index 
)

Return the OptionalArgument from the neural netowrk model that stores intermediate status of solver between epochs by index

Parameters
indexIndex in collection of required OptionalArgument
Returns
the OptionalArgument from the neural netowrk model that stores intermediate status of solver between epochs
Deprecated:
This item will be removed in a future release.
def getSolverOptionalArgumentCollection (   self)

Return the OptionalArgument from the neural netowrk model that stores intermediate status of solver between epochs

Returns
the OptionalArgument from the neural netowrk model that stores intermediate status of solver between epochs
Deprecated:
This item will be removed in a future release.
def getWeightsAndBiases (   self,
  args 
)

Variant 1

Returns table containing all neural network weights and biases

Returns
Table containing all neural network weights and biases
Deprecated:
This item will be removed in a future release.

Variant 2

Returns the weights and biases of the forward layer of neural network as numeric table

Parameters
idxIndex of the backward layer
Returns
Weights and biases derivatives container
Deprecated:
This item will be removed in a future release.
def getWeightsAndBiasesDerivatives (   self,
  args 
)

Variant 1

Returns the weights and biases derivatives of all backward layers of neural network as numeric table

Returns
Weights and biases derivatives container
Deprecated:
This item will be removed in a future release.

Variant 2

Returns the weights and biases derivatives of the backward layer of neural network as numeric table

Parameters
idxIndex of the backward layer
Returns
Weights and biases derivatives container
Deprecated:
This item will be removed in a future release.
def getWeightsAndBiasesStorageStatus (   self)

Returns weights and biases storage status

Returns
Weights and biases storage status. True if weights and biases of all layers stored in one numeric table. False otherwise.
Deprecated:
This item will be removed in a future release.
def initialize_{Float64|Float32} (   self,
  args 
)

Variant 1

Initializes neural network

Parameters
sampleSizeDimensionality of the batch for the input to the first layer
topologyCollection of LayerDescriptor of every inserted layer
parameterParameters of the training
Returns
Status of computations
Deprecated:
This item will be removed in a future release.

Variant 2

Initializes neural network

Parameters
sampleSizeDimensionality of the batch for the input to the first layer
topologyCollection of LayerDescriptor of every inserted layer
parameterParameters of the training
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
Full Names
  • initialize_Float64 is for float64
  • initialize_Float32 is for float32
def serializationTag ( )
def setErrors (   self,
  errors 
)

Sets the error collection to the Model

Parameters
errorsCollection of errors
Deprecated:
This item will be removed in a future release.
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
def setSolverOptionalArgument (   self,
  solverOptionalArgument,
  index 
)

Sets the OptionalArgument to neural netowrk model to store intermediate status of solver between epochs

Parameters
solverOptionalArgumentOptionalArgumentPtr to set in collection
indexIndex in collection of required OptionalArgument
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
def setSolverOptionalArgumentCollection (   self,
  solverOptionalArgumentCollection 
)

Sets the OptionalArgument to neural netowrk model to store intermediate status of solver between epochs

Parameters
solverOptionalArgumentCollectionStructure to store intermediate status of solver
Returns
Status of computations
Deprecated:
This item will be removed in a future release.
def setWeightsAndBiases (   self,
  args 
)

Variant 1

Sets table containing all neural network weights and biases

Parameters
weightsAndBiasesTable containing all neural network weights and biases
Returns
Status of computations
Deprecated:
This item will be removed in a future release.

Variant 2

Sets table containing weights and biases of one forward layer of neural network

Parameters
idxIndex of the forward layer
tableTable containing weights and biases of one forward layer of neural network
Returns
Status of computations
Deprecated:
This item will be removed in a future release.

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

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