Rethinking Calibration of Deep Neural Networks: Do Not Be Afraid of Overconfidence

Open in new window