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

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

Browse the full ICML paper archive.