CogVideoX

Researchers propose SIRUS, a training-free inference-time framework for concept-level unlearning in text-to-video (T2V) models, achieving 70.4% average forgetting success and 25.7% average frame hit o…

Researchers propose LightCrafter, a hybrid pipeline for video relighting that combines physically-based rendering (PBR) with generative video diffusion refinement. By translating a PBR proxy of the in…