"Accelerating Large Batch Training via Gradient Signal to Noise Ratio (GSNR)."

Guoqing Jiang et al. (2023)

Details and statistics

DOI: 10.48550/ARXIV.2309.13681

access: open

type: Informal or Other Publication

metadata version: 2023-09-27

a service of  Schloss Dagstuhl - Leibniz Center for Informatics