DLModel

class mouffet.models.dlmodel.DLModel(opts=None)

Subclass of mouffet.models.model.Model for deep learning models

abstract create_model()

Basic implementation of model creation for deep learning models. By default only creates the network

Returns:

The network created by the “create_net” method

Return type:

Object

abstract load_weights()

Load the weights of the model

Raises:

NotImplementedError – Class must be inherited

save_model(path=None)

Default implementation for deep learning model saving. Calls save_params to save the network options and save_weigths to save the weights of the model.

Parameters:

path ([type], optional) – [description]. Defaults to None.

save_params()

Save network options

abstract save_weights(path=None)

Save the weights of the model

Parameters:

• path (str or pathlib.Path, optional) – The path where the weigths should be saved.

• None. (Defaults to) –

Raises:

NotImplementedError – Class must be inherited

abstract train(training_data, validation_data)

Function call by the training handler to train the model

Parameters:

• training_data (Object) – The training data containing raw data and ground truth

• validation_data (Object) – The validation data containing raw data and ground truth

Raises:

NotImplementedError – Class must be inherited