ADANA thread: log-time weight decay helps AdamW and SOAP but hurts Muon at high overtraining

_katieeverett · x · 2026-09-10

Second thread question: can a similar weight decay schedule improve long-horizon performance? Under fitted coefficient rules, log-time weight decay (Ferbach et al. 2026) helps AdamW, ADANA, and SOAP but harms Muon at high overtraining — the best treatment depends on both optimizer and horizon.

Related event: Momentum-Scheduled ADANA Changes Scaling Exponents on the Overtraining Axis(13 posts)→

Original post →

More from Research

Research channel →