HuggingFace Daily Papers·· 2 天前
FlashVSR+：混合注意力与 Reward-Yield 分布匹配实现高效视频超分
FlashVSR+: Efficient Video Super-Resolution via Hybrid Attention and Reward-Yield Distribution Matching
AI 导读
FlashVSR+ 通过混合注意力架构实现最高 2 倍 DiT 加速和 45% 峰值内存降低，可在单张 24GB RTX 4090 上完成无分块 4K 视频超分。其提出的 Ry-DMD 训练方法用奖励加权教师 rollout 将黑盒奖励纳入分布匹配，无需反传奖励模型，并已推广至 T2I 生成。代码、模型和训练数据已开源。
正文
Abstract:Recent video super-resolution (VSR) models achieve efficient one-step generation through diffusion distillation. However, high-resolution inference remains costly, and distilled models are difficult to align with human preferences. These challenges call for improvements in both model architecture and training strategy. We present \textbf{FlashVSR+} to address both limitations. On the architecture side, we introduce hybrid attention based on the strong locality observed in pretrained VSR models, while retaining sparse global interaction for long-range consistency. It achieves up to $2\times$ DiT speedup and 45\% lower peak memory, enabling spatially untiled 4K VSR on a single 24GB RTX 4090. On the training side, we propose Reward-Yield Distribution Matching Distillation (Ry-DMD) for preference-aligned one-step generation. Ry-DMD uses reward-weighted teacher rollouts to incorporate black-box rewards into distribution matching, without backpropagating through the reward model. We further show that Ry-DMD generalizes beyond VSR to T2I generation. Together, FlashVSR+ delivers stronger restoration quality and substantially improves high-resolution efficiency, with consistent gains in user preference. We release our code, models, and training data at \href{this https URL}{\textcolor{FlashBlue}{\textbf{this http URL}}}.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.17552 [cs.CV] |
| (or arXiv:2610.17552v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.17552 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Ruchang Yao [view email]
[v1]
Tue, 13 Oct 2026 01:22:50 UTC (5,351 KB)
来源：HuggingFace Daily Papers · arxiv.org