"SGD with Clipping is Secretly Estimating the Median Gradient."

Fabian Schaipp et al. (2024)

Details and statistics

DOI: 10.48550/ARXIV.2402.12828

access: open

type: Informal or Other Publication

metadata version: 2024-03-22

a service of  Schloss Dagstuhl - Leibniz Center for Informatics