SwiftVideo: A Unified Framework for Few-Step Video Generation through Trajectory-Distribution Alignment

TL;DR

SwiftVideo is a novel distillation framework that combines trajectory-preserving and distribution-matching strategies to enable high-quality, few-step video generation with reduced computational cost.

Contribution

It introduces continuous-time consistency distillation and dual-perspective alignment to improve stability and performance in few-step video synthesis.

Findings

Outperforms existing methods on OpenVid-1M benchmark

Maintains high video quality with fewer inference steps

Reduces computational overhead significantly

Abstract

Diffusion-based or flow-based models have achieved significant progress in video synthesis but require multiple iterative sampling steps, which incurs substantial computational overhead. While many distillation methods that are solely based on trajectory-preserving or distribution-matching have been developed to accelerate video generation models, these approaches often suffer from performance breakdown or increased artifacts under few-step settings. To address these limitations, we propose \textbf{\emph{SwiftVideo}}, a unified and stable distillation framework that combines the advantages of trajectory-preserving and distribution-matching strategies. Our approach introduces continuous-time consistency distillation to ensure precise preservation of ODE trajectories. Subsequently, we propose a dual-perspective alignment that includes distribution alignment between synthetic and real data along with trajectory alignment across different inference steps. Our method maintains high-quality video generation while substantially reducing the number of inference steps. Quantitative evaluations on the OpenVid-1M benchmark demonstrate that our method significantly outperforms existing approaches in few-step video generation.