twodlearn.templates.supervised module¶

class twodlearn.templates.supervised.AlexNetClassifier(input_shape, n_classes, n_filters, filter_sizes, pool_sizes, n_hidden, options=None, logger_path='tmp', session=None, **kargs)[source]¶

Bases: twodlearn.templates.supervised.LinearClassifier

model[source]¶

class twodlearn.templates.supervised.LinearClassifier(n_inputs=None, n_classes=None, options=None, logger_path='tmp', session=None, **kargs)[source]¶

Bases: twodlearn.templates.supervised.SupervisedEstimator

model[source]¶

monitor[source]¶: Decorator used to specify an optional property inside a model. The decorator works similar to @property, but the specified method correponds to the initialization of the property

train[source]¶

class twodlearn.templates.supervised.MlModel(options=None, logger_path='tmp', session=None)[source]¶

Bases: twodlearn.core.common.TdlProgram

classmethod default_options()[source]¶

property model[source]¶

property monitor[source]¶

property optimizer[source]¶

property options[source]¶

property session[source]¶

property test[source]¶

property train[source]¶

property valid[source]¶

class twodlearn.templates.supervised.MlpClassifier(n_inputs=None, n_classes=None, n_hidden=None, afunction=<function relu>, options=None, logger_path='tmp', session=None, **kargs)[source]¶

Bases: twodlearn.templates.supervised.LinearClassifier

keep_prob[source]¶

model[source]¶

train[source]¶

valid[source]¶

class twodlearn.templates.supervised.Supervised(options=None, tmp_path='tmp')[source]¶

Bases: twodlearn.core.common.TdlProgram

property dataset[source]¶

classmethod default_options()[source]¶

feed_train()[source]¶

feed_valid()[source]¶

property ml_model[source]¶

property options[source]¶

run_training()[source]¶

property tmp_path[source]¶

class twodlearn.templates.supervised.SupervisedEstimator(options=None, logger_path='tmp', session=None, **kargs)[source]¶

Bases: twodlearn.core.common.TdlModel

fit(dataset, max_steps=None, feed_train=None, feed_valid=None)[source]¶

get_save_data()[source]¶

classmethod load(filepath)[source]¶

property logger_path[source]¶

model[source]¶

monitor[source]¶: Decorator used to specify an optional property inside a model. The decorator works similar to @property, but the specified method correponds to the initialization of the property

optimizer[source]¶: Decorator used to specify an optional property inside a model. The decorator works similar to @property, but the specified method correponds to the initialization of the property

predict(inputs)[source]¶

reset_vars()[source]¶: initialize tf variables

save(filepath)[source]¶

session[source]¶

test[source]¶

train[source]¶

valid[source]¶

class twodlearn.templates.supervised.SupervisedMlModel(options=None, logger_path='tmp', session=None)[source]¶

Bases: twodlearn.templates.supervised.MlModel

fit(dataset)[source]¶

class twodlearn.templates.supervised.SupervisedObjective(op, loss, fit_loss, labels)[source]¶: Bases: object

class twodlearn.templates.supervised.SupervisedWrapper(model, inputs, loss, labels, params=None)[source]¶: Bases: object