| title | OPSD-V: On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators |
|---|---|
| authors | Hongyu Liu, Chun Wang, Feng Gao, Xuanhua He, Yue Ma, Ziyu Wan, Yong Zhang, Xiaoming Wei, Qifeng Chen |
| date | 2026-07-09 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2607.08766 |
| affiliation | MeiGen-AI |
| upvotes | 7 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | We propose OPSD-V, an on-policy self-distillation paradigm for post-training few-step autoregressive (AR) video diffusion models. Existing few-step AR video generators can produce long videos with low latency, but still suffer from error accumulation and weakened motion dynamics during long autoregressive rollout. OPSD-V reduces long-horizon degradation while preserving the original few-step inference path. The key idea is to introduce real long-video data as temporal context during training and use it to provide dense trajectory-level supervision. Specifically, the student follows the exact inference-time rollout, generating each chunk conditioned on its own previously generated KV cache. In parallel, the teacher is evaluated at the same student-visited denoising states, but uses a cleaner AR-consistent temporal cache in which older history can be replaced by real-video context. This provides dense denoising-level corrective targets under on-policy AR cache dynamics, without changing the sampler, number of denoising steps, or inference-time cache mechanism. We apply OPSD-V to representative few-step AR video models, including Self-Forcing and LongLive. Experiments show consistent improvements in visual quality, motion dynamics, and VBenchLong scores. A user study with 10 participants comparing 20 video pairs shows that OPSD-V is preferred over the base models in 66.0% of overall-preference judgments (82.5% excluding ties). |
cite
@misc{liu2026opsd,
title={OPSD-V: On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators},
author={Hongyu Liu and Chun Wang and Feng Gao and Xuanhua He and Yue Ma and Ziyu Wan and Yong Zhang and Xiaoming Wei and Qifeng Chen},
year={2026},
eprint={2607.08766},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2607.08766}
}