PulseAugur
EN
LIVE 06:59:55

New Causal-rCM recipe accelerates autoregressive video diffusion

By PulseAugur Editorial · [3 sources] ·

Researchers have introduced Causal-rCM, a novel open recipe for autoregressive video diffusion distillation. This framework unifies teacher-forcing and self-forcing paradigms to enhance streaming video generation and interactive world models. Causal-rCM leverages continuous-time consistency models with a custom FlashAttention-2 kernel, achieving a 10x faster convergence rate than previous methods. The approach has demonstrated state-of-the-art performance in video generation, with a distilled 2-step causal Wan2.1-1.3B model scoring 84.63 on the VBench-T2V benchmark using minimal sampling steps. AI

IMPACT This framework could significantly improve the efficiency and performance of real-time video generation and interactive AI systems.

RANK_REASON The cluster describes a new research paper detailing a novel algorithm and framework for video generation.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 3 sources. How we write summaries →

New Causal-rCM recipe accelerates autoregressive video diffusion

COVERAGE [3]

  1. arXiv cs.LG TIER_1 English(EN) · Kaiwen Zheng, Guande He, Min Zhao, Jintao Zhang, Huayu Chen, Jianfei Chen, Chen-Hsuan Lin, Ming-Yu Liu, Jun Zhu, Qianli Ma ·

    Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models

    arXiv:2606.25473v1 Announce Type: cross 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 d…

  2. arXiv cs.LG TIER_1 English(EN) · Qianli Ma ·

    Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models

    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…

  3. Hugging Face Daily Papers TIER_1 English(EN) ·

    Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models

    Autoregressive video diffusion extends diffusion distillation frameworks to real-time streaming generation through causal training paradigms, achieving state-of-the-art performance with fast convergence and interactive world modeling capabilities.

RELATED ENTITIES

RELATED TOPICS