diffusion-models

Paper Accepted to CVPR 2026: RecTok

Our paper “RecTok: Reconstruction Distillation along Rectified Flow” is accepted to CVPR 2026.

Paper Accepted to ICLR 2026: Vmoba

Our paper “Vmoba: Mixture-of-block attention for video diffusion models” is accepted to ICLR 2026.

Paper Accepted to ICLR 2026: Muddit

Our paper “Muddit: Liberating generation beyond text-to-image with a unified discrete diffusion model” is accepted to ICLR 2026.