通过学习隐空间动态加速图像生成，实现4倍提速且保持高质量。
Accelerating Masked Image Generation by Learning Latent Controlled Dynamics
- 用轻量模型融合历史特征与采样信息，预测特征演化速度场。
- 在Lumina-DiMOO上实现文本到图像生成超4倍加速，质量无损。
- 适合追求高效生成的视觉生成研究者与工业应用开发者。
掩码图像生成模型（MIGMs）虽已取得显著进展，但其效率受限于双向注意力的多步计算。现有方法尝试缓存特征以近似未来特征，但在激进加速下仍存在明显近似误差。我们归因于表达能力不足及未考虑采样信息。为此，提出学习一个轻量模型，融合先前特征与采样标记，回归特征演化的平均速度场。该模型复杂度适中，足以捕捉细微动态，同时远轻于原基模型。将MIGM-Shortcut应用于两种代表性MIGM架构与任务，在当前最先进模型Lumina-DiMOO上，实现文本到图像生成超过4倍加速，且质量保持不变，显著推进了掩码图像生成的帕累托前沿。代码与模型权重已开源。
原文摘要 · Abstract (English)
Masked Image Generation Models (MIGMs) have achieved great success, yet their efficiency is hampered by the multiple steps of bi-directional attention. In fact, there exists notable redundancy in their computation: when sampling discrete tokens, the rich semantics contained in the continuous features are lost. Some existing works attempt to cache the features to approximate future features. However, they exhibit considerable approximation error under aggressive acceleration rates. We attribute this to their limited expressivity and the failure to account for sampling information. To fill this gap, we propose to learn a lightweight model that incorporates both previous features and sampled tokens, and regresses the average velocity field of feature evolution. The model has moderate complexity that suffices to capture the subtle dynamics while keeping lightweight compared to the original base model. We apply our method, MIGM-Shortcut, to two representative MIGM architectures and tasks. In particular, on the state-of-the-art Lumina-DiMOO, it achieves over 4x acceleration of text-to-image generation while maintaining quality, significantly pushing the Pareto frontier of masked image generation. The code and model weights are available at https://github.com/Kaiwen-Zhu/MIGM-Shortcut.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。