Updating AdversarialPatchPytorch to throw an error when user provides labels during an untargeted attack

Author: stekunda

art/attacks/evasion/adversarial_patch/adversarial_patch_pytorch.py

@@ -490,7 +490,7 @@ def generate(  # type: ignore
 
         :param x: An array with the original input images of shape NCHW or input videos of shape NFCHW.
         :param y: True or target labels of format `list[dict[str, Union[np.ndarray, torch.Tensor]]]`, one for each
-                  input image. The fields of the dict are as follows:
+                  input image. For untargeted attacks, this should be None. The fields of the dict are as follows:
 
                   - boxes [N, 4]: the boxes in [x1, y1, x2, y2] format, with 0 <= x1 < x2 <= W and 0 <= y1 < y2 <= H.
                   - labels [N]: the labels for each image.
@@ -515,6 +515,9 @@ def generate(  # type: ignore
 
             y_array: np.ndarray
 
+            if y is not None and not self.targeted:
+                raise ValueError("The untargeted version of AdversarialPatch attack does not use target labels.")
+
             if y is None:  # pragma: no cover
                 logger.info("Setting labels to estimator classification predictions.")
                 y_array = to_categorical(
@@ -533,6 +536,9 @@ def generate(  # type: ignore
             else:
                 self.use_logits = True
         else:
+            if y is not None and not self.targeted:
+                raise ValueError("The untargeted version of AdversarialPatch attack does not use target labels.")
+
             if y is None:  # pragma: no cover
                 logger.info("Setting labels to estimator object detection predictions.")
                 y = self.estimator.predict(x=x)