| title | One-Forcing: Towards Stable One-Step Autoregressive Video Generation |
|---|---|
| authors | Jiaqi Feng, Justin Cui, Yuanhao Ban, Cho-Jui Hsieh |
| date | 2026-05-22 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.23458 |
| affiliation | Tsinghua University |
| upvotes | 4 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Recent advances have substantially improved real-time interactive video generation in the autoregressive regime. However, most existing few-step autoregressive video generation methods, often distilled from a corresponding many-step teacher, default to a 4-step sampling configuration, which still incurs considerable latency during deployment and suffers from severe quality degradation when the number of sampling steps is further reduced, particularly in the one-step setting. Trajectory-style consistency distillation methods often produce videos with weak dynamics, while DMD-based approaches, such as Self-Forcing, tend to yield blurry frames. To address this challenge, we propose One-Forcing, a simple yet effective approach which augments the DMD objective with an auxiliary GAN loss for high-quality and efficient one-step video generation. Experiments on VBench show that One-Forcing achieves a total score of 83.76, establishing state-of-the-art performance among one-step causal video generation methods and remaining competitive with strong many-step approaches. We further demonstrate that one-step framewise autoregressive generation can be achieved stably with merely one-third of the training cost of the chunkwise model, a setting that prior methods have failed to achieve successfully. |
cite
@misc{feng2026one,
title={One-Forcing: Towards Stable One-Step Autoregressive Video Generation},
author={Jiaqi Feng and Justin Cui and Yuanhao Ban and Cho-Jui Hsieh},
year={2026},
eprint={2605.23458},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.23458}
}