4 papers
DyWeight: Dynamic Gradient Weighting for Few-Step Diffusion Sampling
Tong Zhao, Mingkun Lei, Liangyu Yuan +5
Diffusion Models (DMs) have achieved state-of-the-art generative performance across multiple modalities, yet their sampling process remains prohibitively slow due to the need for h…
Fast3Dcache: Training-free 3D Geometry Synthesis Acceleration
Mengyu Yang, Yanming Yang, Chenyi Xu +5
Diffusion models have achieved impressive generative quality across modalities like 2D images, videos, and 3D shapes, but their inference remains computationally expensive due to t…
Taming Video Models for 3D and 4D Generation via Zero-Shot Camera Control
Chenxi Song, Yanming Yang, Tong Zhao +2
Video diffusion models have rich world priors, but their use in spatial tasks is limited by poor control, spatial-temporal inconsistent results, and entangled scene-camera dynamics…
FlowDirector: Training-Free Flow Steering for Precise Text-to-Video Editing
Guangzhao Li, Yanming Yang, Chenxi Song +1
Text-driven video editing aims to modify video content based on natural language instructions. While recent training-free methods have leveraged pretrained diffusion models, they o…