In the code ,

dx = self._alpha*np.power(1 - pt, self._gamma - 1) * (self._gamma * (-1 * pt * pro_) * np.log(pt) + pro_ * (1 - pt)) * 1.0

however, in my inference, they should be
dx = self._alpha*np.power(1 - pt, self._gamma - 1) * (self._gamma * (-1 * pt * pro_) * np.log(pt) + pt * (1 - pt)) * 1.0

Is that right?