def test_neural_network_training():
model = models.NeuralNetworkModel
hyperparams = sample_hp(model.hp_space, rng=RandomState(1))
for dataset in ds.all_datasets:
train, test = dataset.get()
train, test = model.prepare_dataset(train, test, dataset.categorical_features)
estimator = model.build_estimator(hyperparams)
X, y, *_ = train
estimator.fit(X, y)
def test_linear_regression_training():
model = models.LinearRegressionModel
hyperparams = sample_hp(model.hp_space, rng=RandomState(1))
for dataset in ds.all_datasets:
train, test = dataset.get()
train, test = model.prepare_dataset(train, test,
dataset.categorical_features)
estimator = model.build_estimator(hyperparams, test=True)
X, y, *_ = train
estimator.fit(X, y)
def test_gaussian_process_training():
model = models.GaussianProcessModel
hyperparams = sample_hp(model.hp_space, rng=RandomState(1))
for dataset in ds.all_datasets:
try:
print(dataset.__name__)
train, test = dataset.get()
train, test = model.prepare_dataset(train, test,
dataset.categorical_features)
estimator = model.build_estimator(hyperparams)
X, y, *_ = train
estimator.fit(X, y)
except MemoryError:
# This model has high memory requirements and cannot be used on some big datasets
continue
def test_random_forests_hp_space():
sample_hp(models.RandomForestsModel.hp_space)
