Large Learning Rates Improve Generalization: But How Large Are We Talking About?

Ekaterina Lobacheva, Eduard Pokonechny, Maxim Kodryan, Dmitry Vetrov

Keywords: neural network training, large learning rate, generalization

TL;DR: We study the influence of training with different initial learning rates on the generalization.

Abstract: Inspired by recent research that recommends starting neural networks training with large learning rates (LRs) to achieve the best generalization, we explore this hypothesis in detail. Our study clarifies the initial LR ranges that provide optimal results for subsequent training with a small LR or weight averaging. We find that these ranges are in fact significantly narrower than generally assumed. We conduct our main experiments in a simplified setup that allows precise control of the learning rate hyperparameter and validate our key findings in a more practical setting.

Submission Number: 60