Trajectory aware method speeds up video creation with improved quality
TRAC: Trajectory-aware Reuse and Adaptive Correction for Efficient Autoregressive Video Generation
Computer Vision and Pattern Recognition
Summary
Generating videos by predicting one frame after another can be slow and errors can build up over time. The authors introduced TRAC, a method that smartly reuses past information and corrects mistakes to make this process faster and more accurate without needing extra training. They tested it on video datasets and found TRAC gave quicker video generation and better results than existing methods. This approach is helpful for producing high-quality videos efficiently.
What this means in practice
- •For video technology teams: Generate longer high-quality videos faster by reducing error accumulation during sequential frame prediction.
- •For animation software developers: Incorporate efficient autoregressive video generation techniques that maintain structural consistency over time for better animation workflows.
Authors
Jiaxing Song, Weiqi Yan, You Huang, Mingte Qiu, Huazhong Liu, Xiaofeng Zhu, Yunshan Zhong
Abstract
In this paper, we present trajectory-aware reuse and adaptive correction (TRAC), a training-free framework for efficient autoregressive (AR) video generation. Existing acceleration methods mainly target single-trajectory generation with bidirectional attention. AR video generation, by contrast, sequentially couples chunk-level denoising trajectories. Consequently, approximation errors accumulate and propagate through the generation process. TRAC addresses this challenge with three components, including robust cumulative scheduling (RCS), autoregressive trajectory-aware guidance scheduling (ATGS), and spectral structure correction (SSC). RCS selects cache reuse schedules by cumulative rollout error and cross-chunk/prompt variation. ATGS coordinates CFG refreshes along the global AR trajectory. SSC restores low-frequency structure of the first chunk to correct long-term structural loss. Experiments on SkyReels-V2 and FramePack-F1 show that, compared with existing methods, TRAC achieves both the highest inference efficiency and the best generation quality for AR video generation.