Remove check from BoundaryAttack to enable binary classification

Signed-off-by: Beat Buesser <[email protected]>

Author: Beat Buesser

art/attacks/evasion/boundary.py

@@ -133,11 +133,6 @@ def generate(self, x: np.ndarray, y: Optional[np.ndarray] = None, **kwargs) -> n
 
         y = check_and_transform_label_format(y, nb_classes=self.estimator.nb_classes, return_one_hot=False)
 
-        if y is not None and self.estimator.nb_classes == 2 and y.shape[1] == 1:
-            raise ValueError(  # pragma: no cover
-                "This attack has not yet been tested for binary classification with a single output classifier."
-            )
-
         # Get clip_min and clip_max from the classifier or infer them from data
         if self.estimator.clip_values is not None:
             clip_min, clip_max = self.estimator.clip_values