Skip to content
conftrace_
2025 ICML ICML 2025

MERIT: Maximum-normalized Element-wise Ratio for Language Model Large-batch Training

Abstract

No abstract available. No abstract was captured for this paper when its proceedings were scraped - most often an older edition published without one online. How the corpus is built.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Natural Language Processing

Authors

Yang Luo, Zangwei Zheng, Ziheng Qin, Zirui Zhu, Yong Liu, Yang You

Topics

Machine Learning > Optimization & Theory > Neural Network Optimization Natural Language Processing > Resources & Methods > Large Language Models Artificial Intelligence > Core AI > Large Language Models

Keywords

max norm large batch training attention logit lamb merit trust ratio
Download PDF