Fix style check

Signed-off-by: Lei Hsiung <[email protected]>
Trusted-AI · Oct 16, 2023 · cf31929 · cf31929
1 parent 3528185
commit cf31929
Showing 1 changed file with 11 additions and 7 deletions.
diff --git a/art/attacks/evasion/composite_adversarial_attack.py b/art/attacks/evasion/composite_adversarial_attack.py
@@ -214,8 +214,15 @@ def _check_params(self) -> None:
                     _epsilons_range[i][1][0] <= self.epsilons[i][0] <= self.epsilons[i][1] <= _epsilons_range[i][1][1]
                 )
             ):
-                logger.info("The argument `%s` must be an interval within %s of type tuple.", _epsilons_range[i][0], _epsilons_range[i][2])
-                raise ValueError("The argument `{}` must be an interval within {} of type tuple.".format(_epsilons_range[i][0], _epsilons_range[i][2]))
+                logger.info(
+                    "The argument `%s` must be an interval within %s of type tuple.",
+                    _epsilons_range[i][0],
+                    _epsilons_range[i][2],
+                )
+                raise ValueError(
+                    f"The argument `{_epsilons_range[i][0]}` must be an interval "
+                    f"within {_epsilons_range[i][2]} of type tuple."
+                )
 
         if not isinstance(self.early_stop, bool):
             logger.info("The flag `early_stop` has to be of type bool.")
@@ -525,9 +532,7 @@ def caa_linf(
 
         return adv_data, eta
 
-    def update_attack_order(
-        self, images: "torch.Tensor", labels: "torch.Tensor", adv_val: List
-    ) -> None:
+    def update_attack_order(self, images: "torch.Tensor", labels: "torch.Tensor", adv_val: List) -> None:
         """
         Update the specified attack ordering.
         :param images: A tensor of a batch of original inputs to be attacked.
@@ -600,7 +605,6 @@ def caa_attack(self, images: "torch.Tensor", labels: "torch.Tensor") -> "torch.T
         """
         import torch
 
-        attack = self.attack_dict
         adv_img = images.detach().clone()
         adv_val_saved = torch.zeros((self.seq_num, self.batch_size), device=self.device)
         adv_val = [self.adv_val_space[idx] for idx in range(self.seq_num)]
@@ -623,7 +627,7 @@ def caa_attack(self, images: "torch.Tensor", labels: "torch.Tensor") -> "torch.T
 
             for tdx in range(self.seq_num):
                 idx = self.curr_seq[tdx]
-                adv_img, adv_val_updated = attack[idx](adv_img, adv_val[idx], labels)
+                adv_img, adv_val_updated = self.attack_dict[idx](adv_img, adv_val[idx], labels)
                 if idx != self.linf_idx:
                     adv_val[idx] = adv_val_updated