Nabu-asr
Public Member Functions | Public Attributes | List of all members
nabu.neuralnetworks.models.model.Model Class Reference

a general class for an encoder decoder system More...

Inheritance diagram for nabu.neuralnetworks.models.model.Model:

Public Member Functions

def __init__ (self, conf, trainlabels, constraint)
 Model constructor. More...
 
def __call__ (self, inputs, input_seq_length, targets, target_seq_length, is_training)
 Add the neural net variables and operations to the graph. More...
 
def variables (self)
 get a list of the models's variables
 

Public Attributes

 conf
 
 input_names
 
 output_names
 
 output_dims
 
 encoder
 
 decoder
 

Detailed Description

a general class for an encoder decoder system

Constructor & Destructor Documentation

§ __init__()

def nabu.neuralnetworks.models.model.Model.__init__ (   self,
  conf,
  trainlabels,
  constraint 
)

Model constructor.

Parameters
confThe model configuration as a configparser object
trainlabelsthe number of extra labels required by the trainer
constrainta constraint for the model paramaters

Member Function Documentation

§ __call__()

def nabu.neuralnetworks.models.model.Model.__call__ (   self,
  inputs,
  input_seq_length,
  targets,
  target_seq_length,
  is_training 
)

Add the neural net variables and operations to the graph.

Parameters
inputsthe inputs to the neural network, this is a dictionary of [batch_size x time x ...] tensors
input_seq_lengthThe sequence lengths of the input utterances, this is a dictionary of [batch_size] vectors
targetsthe targets to the neural network, this is a dictionary of [batch_size x time x ...] tensors.
target_seq_lengthThe sequence lengths of the target utterances, this is a dictionary of [batch_size] vectors
is_trainingwhether or not the network is in training mode
Returns
  • output logits, which is a dictionary of [batch_size x time x ...]
tensors
  • the output logits sequence lengths which is a dictionary of [batch_size] vectors

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