keras_model_to_pmml module

Classes used in

class keras_model_to_pmml.KerasDataDictionary(dataSet, predictedClasses)[source]

Bases: PMML43Ext.DataDictionary

KerasDataDictionary stores the class information to be predicted in the PMML model. The current implementation takes care of the Imagenet class label by giving dataset name as dataSet parameter.

  • dataSet (String) – Name of the dataset
  • predictedClasses (List) – List of class names or values to be predicted.

Return type:

Nyoka’s Dictionary Object

class keras_model_to_pmml.KerasHeader(description, copyright)[source]

Bases: PMML43Ext.Header

Creates header for Keras PMML model file using Nyoka

  • copyright (String) – Adds the information about the copyright.
  • description (String) – Description of the PMML file provided as a default
  • Timestamp (Datetime) – Timestamp of the time when the file is created

Return type:

Nyoka header object

class keras_model_to_pmml.KerasLocalTransformations(model_name)[source]

Bases: PMML43Ext.LocalTransformations

KerasLocalTransformations provides the information about the list of transformations applied to the data.

Parameters:model_name (String) – Name of the model (internally used to be specific for Keras)
Return type:Nyoka’s Transformations Object
class keras_model_to_pmml.KerasMiningSchema(dataSet=None)[source]

Bases: PMML43Ext.MiningSchema

KerasMiningSchema stores the attributes which are used to build the model.

Parameters:dataSet (String) – Name of the dataset
Return type:Nyoka’s Mining Schema Object
class keras_model_to_pmml.KerasNetwork(keras_model, model_name, dataSet=None, predictedClasses=None)[source]

Bases: PMML43Ext.DeepNetwork

KerasNetwork creates the DeepNetwork object which stores the NetworkLayer in sequence to define the architecture.

  • model_name (String) – Name of the model
  • functionName (String) – Regression or Classification, currently supports classification functionName
  • numberOfLayers (Int) – Number of layers in the architecture
  • isScorable (Boolean) – True or False
  • Extension (Nyoka's extention tag) – Allows to pass extra information in Nyoka objects
  • MiningSchema (Nyoka's Mining schema object) – Nyoka’s miningschema object to be passed
  • Output (Nyoka's Output object) – Nyoka’s Output object to be passed
  • LocalTransformations (Nyoka's LocalTransformations object) – Nyoka’s LocalTransformations object to be passed
  • NetworkLayer (Nyoka's LocalTransformations object) – Nyoka’s NetworkLayer object to be passed

Return type:

Nyoka’s DeepNetwork Object

class keras_model_to_pmml.KerasNetworkLayer(layer, layer_type)[source]

Bases: PMML43Ext.NetworkLayer

Creates Networklayer of PMML which has information about the layer type, weight matrix and bias matrix and their properties.

  • inputFieldName (String) – This parameter is required only for Input layer in keras
  • layerType (String) – Any Keras layer (e.g. Input, Dense, Conv2D)
  • connectionLayerId (String) – Name of the previous layer ID
  • layerId (String) – Layer ID for defined layer
  • normalizationMethod (String) – Name of normalization method here
  • LayerParameters (Nyoka LayerParamter Object) – Nyoka’s LayerParameter object which has information of Layerparamters (eg, input dimension and output dimension).
  • LayerWeights (Nyoka's LayerWeights object) – LayerWeights goes inside the LayerParameters object and provide information about the weigth matrix of the layer.
  • LayerBias (Nyoka's LayerBias object) – LayerBias goes inside the LayerParameters object and provide value of the bias matrix.

Return type:

Nyoka NetworkLayer object

class keras_model_to_pmml.KerasOutput(predictedClasses=None)[source]

Bases: PMML43Ext.Output

KerasOutput provides the information about the output representation of the PMML. (e.g. Predicted classes, probabilities)

Parameters:predictedClasses (List) – List of Classes for which model has been trained
Return type:Nyoka’s Output Object
class keras_model_to_pmml.KerasToPmml(keras_model, model_name='KerasNet', description='Keras Models in PMML', copyright='Internal User', dataSet=None, predictedClasses=None)[source]

Bases: PMML43Ext.PMML

KerasToPmml exports the Keras model object into PMML file using nyoka.

  • keras_model (keras model object) – Keras model object
  • model_name (String) – Name to be given to the model in PMML.
  • dataSet (String (Optional)) – Name of the dataset
  • predictedClasses (List) – List of the class names for which model has been trained

Return type:

Creates PMML object, this can be saved in file using export function