arXiv:2508.10963v2 Announce Type: replace 
Abstract: Visual generation includes both image and video generation, training probabilistic models to create coherent, diverse, and semantically faithful content from scratch. While early research focused on unconditional sampling, practitioners now demand controllable generation that allows precise specification of layout, pose, motion, or style. While ControlNet grants precise spatial-temporal control, its auxiliary branch markedly increases latency and introduces redundant computation in both uncontrolled regions and denoising steps, especially for video. To address this problem, we introduce EVCtrl, a lightweight, plug-and-play control adapter that slashes overhead without retraining the model. Specifically, we propose a spatio-temporal dual caching strategy for sparse control information. For spatial redundancy, we first profile how each layer of DiT-ControlNet responds to fine-grained control, then partition the network into global and local functional zones. A locality-aware cache focuses computation on the local zones that truly need the control signal, skipping the bulk of redundant computation in global regions. For temporal redundancy, we selectively omit unnecessary denoising steps to improve efficiency. Extensive experiments on CogVideo-Controlnet, Wan2.1-Controlnet, and Flux demonstrate that our method is effective in image and video control generation without the need for training. For example, it achieves 2.16 and 2.05 times speedups on CogVideo-Controlnet and Wan2.1-Controlnet, respectively, with almost no degradation in generation quality.Codes are available in the supplementary materials.

تم تقديم EVCtrl كأداة تحكم فعالة لتوليد الصور، حيث تلبي الحاجة إلى توليد الصور والفيديوهات القابلة للتحكم دون الحاجة إلى إعادة تدريب النماذج. تستخدم هذه الابتكار استراتيجية تخزين مؤقت مزدوجة زمنية ومكانية لتحسين الأداء، خاصة في السيناريوهات التي تتطلب تحكمًا محدودًا.

EVCtrl se ha presentado como un adaptador de control eficiente para la generación visual, abordando la necesidad de generación de imágenes y videos controlables sin la sobrecarga de reentrenar modelos. Esta innovación utiliza una estrategia de almacenamiento en caché dual espaciotemporal para optimizar el rendimiento, especialmente en escenarios de control escaso.

EVCtrl a été introduit comme un adaptateur de contrôle efficace pour la génération visuelle, répondant au besoin de génération d'images et de vidéos contrôlables sans le surcoût de la réentraînement des modèles. Cette innovation utilise une stratégie de mise en cache duale spatio-temporelle pour optimiser les performances, en particulier dans les scénarios de contrôle sparse.

EVCtrl has been introduced as an efficient control adapter for visual generation, addressing the need for controllable image and video generation without the overhead of retraining models. This innovation utilizes a spatio-temporal dual caching strategy to optimize performance, particularly in sparse control scenarios.

EVCtrl: Efficient Control Adapter for Visual Generation

Was this article worth reading? Share it

Republiclabs.ai

Glima

Videolulu