7 papers · 1 filter
LiveEdit: Towards Real-Time Diffusion-Based Streaming Video Editing
Xinyu Wang, Chongbo Zhao, Fangneng Zhan +1
Streaming video editing has made rapid progress, yet practical deployment is still limited by two core issues: maintaining stable backgrounds and non-edited regions over time, and…
EasyVFX: Frequency-Driven Decoupling for Resource-Efficient VFX Generation
Yue Ma, Xu Ye, Qinghe Wang +9
Generating high-fidelity visual effects (VFX) typically demands massive datasets and prohibitive computational power due to the intricate coupling of spatial textures and temporal…
Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion
Peiliang Cai, Evelyn Zhang, Jiacheng Liu +8
Recent advances in autoregressive video diffusion have enabled sequential and streaming video generation. However, long-horizon generation requires increasingly large KV caches, ma…
DiTraj: training-free trajectory control for video diffusion transformer
Cheng Lei, Jiayu Zhang, Yue Ma +6
Diffusion Transformers (DiT)-based video generation models with 3D full attention exhibit strong generative capabilities. Trajectory control represents a user-friendly task in the…
Forecast then Calibrate: Feature Caching as ODE for Efficient Diffusion Transformers
Shikang Zheng, Liang Feng, Xinyu Wang +8
Diffusion Transformers (DiTs) have demonstrated exceptional performance in high-fidelity image and video generation. To reduce their substantial computational costs, feature cachin…
HERO: Hierarchical Extrapolation and Refresh for Efficient World Models
Quanjian Song, Xinyu Wang, Donghao Zhou +3
Generation-driven world models create immersive virtual environments but suffer slow inference due to the iterative nature of diffusion models. While recent advances have improved…