V*Bench

'VBench' is a benchmark used to evaluate the performance of sparse attention mechanisms in video generation tasks.

Papers using V*Bench (7)

LearniBridge: Learnable Calibration of Feature Caching for Diffusion Models Acceleration2026Training-free sparse attention based on cumulative energy filtering2026Light Forcing: Accelerating Autoregressive Video Diffusion via Sparse Attention2026Towards One-step Causal Video Generation via Adversarial Self-Distillation2025Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders2025VideoPainter: Any-length Video Inpainting and Editing with Plug-and-Play Context Control2025Goku: Flow Based Video Generative Foundation Models2025