def test_lenet_specific_input_shape_classes(self):
size = (1, 18, 18)
classes = 3
data = np.ones((1, ) + size)
input_data = torch.Tensor(data)
lenet = LeNet(size, classes=classes)
output_shape = lenet(input_data).detach().numpy().shape
self.assertEqual(output_shape, (1, classes))
def get_estimator(epsilon=0.04,
epochs=10,
batch_size=32,
max_train_steps_per_epoch=None,
max_eval_steps_per_epoch=None,
save_dir=tempfile.mkdtemp()):
# step 1
train_data, eval_data = cifair10.load_data()
test_data = eval_data.split(0.5)
pipeline = fe.Pipeline(train_data=train_data,
eval_data=eval_data,
test_data=test_data,
batch_size=batch_size,
ops=[
Normalize(inputs="x",
outputs="x",
mean=(0.4914, 0.4822, 0.4465),
std=(0.2471, 0.2435, 0.2616)),
ChannelTranspose(inputs="x", outputs="x"),
])
# step 2
model = fe.build(model_fn=lambda: LeNet(input_shape=(3, 32, 32)),
optimizer_fn="adam")
network = fe.Network(ops=[
Watch(inputs="x"),
ModelOp(model=model, inputs="x", outputs="y_pred"),
CrossEntropy(inputs=("y_pred", "y"), outputs="base_ce"),
FGSM(data="x", loss="base_ce", outputs="x_adverse", epsilon=epsilon),
ModelOp(model=model, inputs="x_adverse", outputs="y_pred_adv"),
CrossEntropy(inputs=("y_pred_adv", "y"), outputs="adv_ce"),
Average(inputs=("base_ce", "adv_ce"), outputs="avg_ce"),
UpdateOp(model=model, loss_name="avg_ce")
])
# step 3
traces = [
Accuracy(true_key="y", pred_key="y_pred", output_name="base accuracy"),
Accuracy(true_key="y",
pred_key="y_pred_adv",
output_name="adversarial accuracy"),
BestModelSaver(model=model,
save_dir=save_dir,
metric="adv_ce",
save_best_mode="min"),
]
estimator = fe.Estimator(
pipeline=pipeline,
network=network,
epochs=epochs,
traces=traces,
max_train_steps_per_epoch=max_train_steps_per_epoch,
max_eval_steps_per_epoch=max_eval_steps_per_epoch,
monitor_names=["base_ce", "adv_ce"])
return estimator
def build_model(self):
"""
Define the FastEstimator model architecture.
Args:
None
Returns:
model: Union[tf.keras.sequential, nn.module]
"""
model = fe.build(model_fn=lambda: LeNet(input_shape=(3, 32, 32)),
optimizer_fn="adam",
model_name="adv_model")
return model
def test_check_input_shape(self):
with self.subTest("length not 3"):
with self.assertRaises(ValueError):
LeNet._check_input_shape((1, ))
with self.subTest("width or height is smaller than 18"):
with self.assertRaises(ValueError):
LeNet._check_input_shape((1, 13, 18))
with self.assertRaises(ValueError):
LeNet._check_input_shape((1, 18, 1))
with self.subTest("both are not smaller than 18"):
LeNet._check_input_shape((1, 18, 18))
LeNet._check_input_shape((32, 18, 100))
LeNet._check_input_shape((64, 200, 18))
def test_lenet_default(self):
data = np.ones((1, 1, 28, 28))
input_data = torch.Tensor(data)
lenet = LeNet()
output_shape = lenet(input_data).detach().numpy().shape
self.assertEqual(output_shape, (1, 10))