AI RESEARCH
Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generation
arXiv CS.CV
•
ArXi:2604.03118v1 Announce Type: new Distilling video generation models to extremely low inference budgets (e.g., 2--4 NFEs) is crucial for real-time deployment, yet remains challenging. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion. Distribution matching distillation (DMD) can recover sharp, mode-seeking samples, but its local