Paper Detail
Kaiwen Zheng, Guande He, Min Zhao, Jintao Zhang, Huayu Chen, Jianfei Chen, Chen-Hsuan Lin, Ming-Yu Liu, Jun Zhu, Qianli Ma
Autoregressive video diffusion with causal diffusion transformers has emerged as a major paradigm for real-time streaming video generation and action-conditioned interactive world models. In this work, we extend rCM, an advanced diffusion distillation framework, to autoregressive video diffusion. The core philosophy of rCM lies in the complementarity between forward and reverse divergences, represented by consistency models (CMs) and distribution matching distillation (DMD), respectively, in diffusion distillation. This philosophy naturally carries over to the autoregressive setting, where teacher-forcing (TF) provides an offline, forward-divergence causal training paradigm, while self-forcing (SF) corresponds to an on-policy, reverse-divergence refinement. Our contributions are: (1) through extensive experiments, we show that teacher-forcing CM is currently the best complement to self-forcing DMD as an initialization strategy (2) we present the first implementation of teacher-forcing-based continuous-time CMs (e.g., sCM/MeanFlow) for autoregressive video diffusion, enabled by our custom-mask FlashAttention-2 JVP kernel, achieving 10times faster convergence compared to discrete-time CMs (dCMs) (3) we introduce Causal-rCM, a leading, unified, and scalable algorithm-infrastructure open recipe for diffusion distillation and causal training (4) we achieve state-of-the-art streaming video generation performance in both frame-wise and chunk-wise settings, using only synthetic data for training. Notably, our distilled 2-step causal Wan2.1-1.3B model achieves a VBench-T2V score of 84.63 with only 1 or 2 sampling steps. We further apply Causal-rCM to Cosmos 3, an advanced omnimodal world foundation model for physical AI with action-conditioned generation capability, enabling an interactive world model.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{zheng2026causal,
title = {Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models},
author = {Kaiwen Zheng and Guande He and Min Zhao and Jintao Zhang and Huayu Chen and Jianfei Chen and Chen-Hsuan Lin and Ming-Yu Liu and Jun Zhu and Qianli Ma},
year = {2026},
abstract = {Autoregressive video diffusion with causal diffusion transformers has emerged as a major paradigm for real-time streaming video generation and action-conditioned interactive world models. In this work, we extend rCM, an advanced diffusion distillation framework, to autoregressive video diffusion. The core philosophy of rCM lies in the complementarity between forward and reverse divergences, represented by consistency models (CMs) and distribution matching distillation (DMD), respectively, in dif},
url = {https://huggingface.co/papers/2606.25473},
keywords = {autoregressive video diffusion, causal diffusion transformers, diffusion distillation, consistency models, distribution matching distillation, teacher-forcing, self-forcing, continuous-time CMs, discrete-time CMs, FlashAttention-2, JVP kernel, Causal-rCM, VBench-T2V, omnimodal world foundation model, action-conditioned generation, huggingface daily},
eprint = {2606.25473},
archiveprefix = {arXiv},
}
{}