Toward Optimal Run Racing: Application to Deep Learning Calibration

   Abstract