autoBOTLib.learning package

Submodules

autoBOTLib.learning.hyperparameter_configurations module

autoBOTLib.learning.scikit_based module

autoBOTLib.learning.scikit_based.scikit_learners(final_run, tmp_feature_space, train_targets, learner_hyperparameters, learner_preset, learner, task, scoring_metric, n_fold_cv, validation_percentage, random_seed, verbose, validation_type, num_cpu)

An auxilliary method which conducts sklearn-based learning

autoBOTLib.learning.torch_sparse_nn module

class autoBOTLib.learning.torch_sparse_nn.E2EDatasetLoader(features, targets=None)

Bases: Generic[torch.utils.data.dataset.T_co]

A generic toch dataframe loader adapted for sparse (CSR) matrices.

__init__(features, targets=None)

Initialize self. See help(type(self)) for accurate signature.

autoBOTLib.learning.torch_sparse_nn.to_one_hot(lbx)

class autoBOTLib.learning.torch_sparse_nn.HyperParamNeuralObject(hyperparam_space, verbose=0, device='cpu', metric='f1_macro')

Bases: object

Meta learning object governing hyperparameter optimization

__init__(hyperparam_space, verbose=0, device='cpu', metric='f1_macro')

Initialize self. See help(type(self)) for accurate signature.

fit(X, Y, refit=False, n_configs=1)

A generic fit method, resembling GridSearchCV

class autoBOTLib.learning.torch_sparse_nn.GenericFFNN(input_size, num_classes, hidden_layer_size, num_hidden=2, dropout=0.02, device='cuda')

Bases: torch.nn.modules.module.Module

__init__(input_size, num_classes, hidden_layer_size, num_hidden=2, dropout=0.02, device='cuda')

Initializes internal Module state, shared by both nn.Module and ScriptModule.

hadamard_act(x)

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_importances()

training: bool

class autoBOTLib.learning.torch_sparse_nn.SFNN(batch_size=32, num_epochs=32, learning_rate=0.001, stopping_crit=10, hidden_layer_size=64, dropout=0.2, num_hidden=2, device='cpu', verbose=0, *args, **kwargs)

Bases: object

__init__(batch_size=32, num_epochs=32, learning_rate=0.001, stopping_crit=10, hidden_layer_size=64, dropout=0.2, num_hidden=2, device='cpu', verbose=0, *args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

fit(features, labels)

predict(features, return_proba=False)

get_importances(features)

predict_proba(features)

autoBOTLib.learning.torch_sparse_nn.cross_val_score_nn(config, X, Y, metric='f1_macro')

A method which performs cross-validation and returns top score

autoBOTLib.learning.torch_sparse_nn.get_random_config(cdict)

Select a random hyperparam configuration

autoBOTLib.learning.torch_sparse_nn.hyper_opt_neural(X, Y, refit=True, verbose=1, device='cpu', learner_preset='default', metric='f1_macro')

Generic hyperoptimization routine

autoBOTLib.learning.torch_sparse_nn.torch_learners(final_run, X, Y, custom_hyperparameters, learner_preset, learner, task, metric, num_folds, validation_percentage, random_seed, verbose, validation_type, num_cpu, device='cpu')

A method for searching the architecture space

Module contents