FRUGAL: Memory-Efficient Optimization by Reducing State Overhead for Scalable Training

Documenti analoghi

• LionMuon: Alternating Spectral and Sign Descent for Efficient Training
  di: Bolatov, Arman, et al.
  Pubblicazione: (2026)
• Random-reshuffled SARAH does not need a full gradient computations
  di: Beznosikov, Aleksandr, et al.
  Pubblicazione: (2021)
• Label Privacy in Split Learning for Large Models with Parameter-Efficient Training
  di: Zmushko, Philip, et al.
  Pubblicazione: (2024)
• Just a Simple Transformation is Enough for Data Protection in Vertical Federated Learning
  di: Semenov, Andrei, et al.
  Pubblicazione: (2024)
• Preconditioned Norms: A Unified Framework for Steepest Descent, Quasi-Newton and Adaptive Methods
  di: Veprikov, Andrey, et al.
  Pubblicazione: (2025)