SHARPNESS-AWARE MINIMIZATION FOR EFFICIENTLY IMPROVING GENERALIZATIONASAM: Adaptive Sharpness-Aware Minimization for Scale-Invariant Learning of Deep Neural
