Causal Forcing++: Scalable Few-Step Autoregressive Diffusion Distillation for Real-Time Interactive Video Generation

Min Zhao, Hongzhou Zhu, Kaiwen Zheng, Zihan Zhou, Bokai Yan, Xinyuan Li, Xiao Yang, Chongxuan Li, Jun Zhu

arXiv:2605.15141 · 2026-05-13 공개 · arXiv · PDF

video-generation autoregressive-models few-step-generation vbench causal-forcing causal-consistency diffusion-distillation real-time-interactive

Abstract

Real-time interactive video generation requires low-latency, streaming, and controllable rollout. Existing autoregressive (AR) diffusion distillation methods have achieved strong results in the chunk-wise 4-step regime by distilling bidirectional base models into few-step AR students, but they remain limited by coarse response granularity and non-negligible sampling latency. In this paper, we study a more aggressive setting: frame-wise autoregression with only 1--2 sampling steps. In this regime, we identify the initialization of a few-step AR student as the key bottleneck: existing strategies are either target-misaligned, incapable of few-step generation, or too costly to scale. We propose \textbf{Causal Forcing++}, a principled and scalable pipeline that uses \emph{causal consistency distillation} (causal CD) for few-step AR initialization. The core idea is that causal CD learns the same AR-conditional flow map as causal ODE distillation, but obtains supervision from a single online teacher ODE step between adjacent timesteps, avoiding the need to precompute and store full PF-ODE trajectories. This makes the initialization both more efficient and easier to optimize. The resulting pipeline, \ours, surpasses the SOTA 4-step chunk-wise Causal Forcing under the \textit{\textbf{frame-wise 2-step setting}} by 0.1 in VBench Total, 0.3 in VBench Quality, and 0.335 in VisionReward, while reducing first-frame latency by 50\% and Stage 2 training cost by $\sim$$4\times$. We further extend the pipeline to action-conditioned world model generation in the spirit of Genie3. Project Page: https://github.com/thu-ml/Causal-Forcing and https://github.com/shengshu-ai/minWM .

한국어 요약

📋 한 줄 요약

**[비디오 생성 / 확산 모델 증류]** 실시간 상호작용 비디오 생성을 위해 프레임 단위 1~2 스텝 자기회귀 확산 증류를 가능케 하는 인과 일관성 증류(causal CD) 기반 파이프라인 Causal Forcing++ 제안.

🎯 핵심 기여도

💡 핵심 아이디어

완전한 PF-ODE 궤적을 사전 저장·학습에 사용하는 ODE distillation 대신, 인접 timestep 간의 단일 teacher 스텝에서 얻은 동등한 supervision을 사용하면 동일한 AR-conditional flow map을 더 효율적으로 학습할 수 있다는 통찰.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 게임/시뮬레이션/인터랙티브 미디어에서 요구되는 저지연 스트리밍 비디오 생성에 필요한 frame-wise few-step 체제를 실제로 가능케 함. **한계**: 1~2 스텝 체제의 안정성은 teacher 품질에 강하게 의존하며, 매우 긴 시계열에서의 누적 오차 영향은 추가 검증 필요.

🚀 실용적 활용