f9dd2e3636db61ee1b4d32ee4f494fd9e71adc12,cleverhans/attacks.py,FastGradientMethod,generate_np,#FastGradientMethod#Any#Any#,159

Before Change

        
        Generate adversarial samples and return them in a Numpy array.
        
        if self.default_graph is None:
            error_string = "The attack symbolic graph was not generated."
            raise Exception(error_string)
        if self.nb_calls_generate > 1:
            warnings.warn("Attack was generated symbolically multiple "
                          "times, using graph defined by first call.")

        // Verify label placeholder was defined previously if using true labels
        if params["Y"] is not None:
            error = "True labels given but label placeholder missing in _init_"
            assert self.y is not None, error

After Change

        import tensorflow as tf

        // Generate this attack"s graph if it hasn"t been done previously
        if not hasattr(self, "_x"):
            input_shape = list(X.shape)
            input_shape[0] = None
            self._x = tf.placeholder(tf.float32, shape=input_shape)
            self._x_adv = self.generate(self._x)

        // Run symbolic graph without or with true labels
        if params["Y"] is None:
            feed_dict = {self._x: X}
        else:
            feed_dict = {self._x: X, self.y: params["Y"]}

Instances