def test_adversarial_training_range():
"""
This unit test checks the input range check in adversarial_training.py.
"""
model = torchensemble.AdversarialTrainingClassifier(estimator=MLP_clf,
n_estimators=2,
cuda=False)
model.set_optimizer("Adam")
# Prepare data
train = TensorDataset(X_train, y_train_clf)
train_loader = DataLoader(train, batch_size=2)
# Training
with pytest.raises(ValueError) as excinfo:
model.fit(train_loader)
assert "input range of samples passed to adversarial" in str(excinfo.value)
def test_adversarial_training():
model = torchensemble.AdversarialTrainingClassifier(estimator=MLP,
n_estimators=2,
cuda=False)
# Epochs
with pytest.raises(ValueError) as excinfo:
model.fit(train_loader, epochs=-1)
assert "number of training epochs" in str(excinfo.value)
# Epsilon
with pytest.raises(ValueError) as excinfo:
model.fit(train_loader, epsilon=2)
assert "step used to generate adversarial samples" in str(excinfo.value)
# Log interval
with pytest.raises(ValueError) as excinfo:
model.fit(train_loader, log_interval=-1)
assert "number of batches to wait" in str(excinfo.value)