optimizers · apple

AdEMAMix

Adam variant that keeps two exponential moving averages of the gradient at different decay rates instead of one, so the optimiser can react to recent gradients while still retaining signal from much older ones. The stated motivation is that a single EMA cannot be both responsive and long-memoried, and that older gradients stay informative for longer than one decay rate can express.

Verify

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.
Re-check this yourself → Source → Full record →

What is recorded

Publisherapple
OriginUnited States
Licenceresearch-paper

Elsewhere in the registry

All optimizers → Browse with filters → What will run on my hardware? →

Registry record as of 2026-08-14. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.