arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2102.06356cs.LGstat.ML
收藏

一个大批量优化器现实检查:传统的,通用的优化器足以跨越批量大小

A Large Batch Optimizer Reality Check: Traditional, Generic Optimizers Suffice Across Batch Sizes

Google Research(谷歌研究院)

Zachary Nado, Justin M. Gilmer, Christopher J. Shallue, Rohan Anil, George E. Dahl

摘要:

最近,LARS和LAMB优化器已经被提出用于使用大批量更快地训练神经网络。LARS和LAMB分别为Heavy-ball momentum和Adam的更新规则添加了逐层规范化,并在著名的基准测试和深度学习库中流行。然而,如果没有与标准优化器进行公平的比较,LARS和LAMB是否比传统的通用算法有任何好处仍然是一个悬而未决的问题。在这项工作中,我们证明了标准的优化算法,如Nesterov动量和亚当可以匹配或超过LARS和LAMB在大批量的结果。我们的研究结果建立了新的,更强大的基线,为未来的比较,在这些批量大小,并揭示了比较优化的神经网络训练更普遍的困难。

英文摘要:

Recently the LARS and LAMB optimizers have been proposed for training neural networks faster using large batch sizes. LARS and LAMB add layer-wise normalization to the update rules of Heavy-ball momentum and Adam, respectively, and have become popular in prominent benchmarks and deep learning libraries. However, without fair comparisons to standard optimizers, it remains an open question whether LARS and LAMB have any benefit over traditional, generic algorithms. In this work we demonstrate that standard optimization algorithms such as Nesterov momentum and Adam can match or exceed the results of LARS and LAMB at large batch sizes. Our results establish new, stronger baselines for future comparisons at these batch sizes and shed light on the difficulties of comparing optimizers for neural network training more generally.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑