support for nondeterministic input shape to a model by gradient matching attack.

Signed-off-by: Taesung Lee <[email protected]>

Author: TS-Lee

art/attacks/poisoning/gradient_matching_attack.py

@@ -151,8 +151,8 @@ def __initialize_poison_tensorflow(
 
         :param x_trigger: A list of samples to use as triggers.
         :param y_trigger: A list of target classes to classify the triggers into.
-        :param x_train: A list of training data to poison a portion of.
-        :param y_train: A list of labels for x_train.
+        :param x_poison: A list of training data to poison a portion of.
+        :param y_poison: A list of true labels for x_poison.
         """
         # pylint: disable=no-name-in-module
         from tensorflow.keras import backend as K