Video Diffusion Models
PulseAugur coverage of Video Diffusion Models — every cluster mentioning Video Diffusion Models across labs, papers, and developer communities, ranked by signal.
2 天有情绪数据
-
New PREX framework enables faithful 4D video editing
Researchers have developed PREX, a novel framework for faithful 4D video editing that addresses the challenge of preserving original regions while synthesizing new content. The method identifies and corrects an "Evidenc…
-
New methods boost video diffusion model efficiency and quality
Researchers have developed several new techniques to improve video diffusion models, focusing on efficiency and quality. One approach, LocalDPO, optimizes alignment at a localized spatio-temporal region level for better…
-
New Diffusion-APO method aligns video diffusion models with user intent
Researchers have introduced Diffusion-APO, a new method for aligning video diffusion models with human preferences. This approach addresses the gap between training noise distributions and real-world inference by synchr…
-
New research explores 3D consistency, LoRA transferability, and unified frameworks for video diffusion models.
Researchers have developed new methods to improve video generation using diffusion models. One approach, Geometry Forcing, integrates 3D representations with video diffusion models to enhance geometric consistency and v…