arXiv - CSCL: "Two-Tailed Averaging: Anytime, Adaptive, Once-in-…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

Two-Tailed Averaging: Anytime, Adaptive, Once-in-a-While Optimal Weight Averaging for Better Generalization. (arXiv:2209.12581v3 [stat.ML] UPDATED)

http://arxiv.org/abs/2209.12581 #arXiv #NLProc

Apr 18, 2023, 03:07 · · arxiv-cscl · · ·

Sign in to participate in the conversation