Large Catapults in Momentum Gradient Descent with Warmup: An Empirical Study

Abstract

Although gradient descent with momentum is widely used in modern deep learning, a concrete understanding of its effects on the training trajectory still remains elusive. In this work, we empirically show that momentum gradient descent with a large learning rate and learning rate warmup displays large catapults, driving the iterates towards flatter minima than those found by gradient descent. We then provide empirical evidence and theoretical intuition that the large catapult is caused by momentum “amplifying” the self-stabilization effect (Damian et al., 2023).

Chulhee Yun

Assistant Professor

I am an assistant professor at KAIST AI. I am interested in optimization and machine learning theory.