TensorX
返回文献探索

Paper · arXiv 2602.06028

Context Forcing: Consistent Autoregressive Video Generation with Long Context

Shuo Chen, Cong Wei, Sun Sun, Ping Nie, Kai Zhou, Ge Zhang, Ming-Hsuan Yang, Wenhu Chen

36 upvotesFebruary 5, 2026arXiv 预印本
AI 摘要

Context Forcing addresses student-teacher mismatch in long video generation by using a long-context teacher to guide long-rollout students through a Slow-Fast Memory architecture that extends context length beyond 20 seconds.

streaming tuning strategieslong-context studentshort-context teacherstudent-teacher mismatchcontext management systemSlow-Fast Memorylong-term consistencylong video generation

Abstract

Recent approaches to real-time long video generation typically employ streaming tuning strategies, attempting to train a long-context student using a short-context (memoryless) teacher. In these frameworks, the student performs long rollouts but receives supervision from a teacher limited to short 5-second windows. This structural discrepancy creates a critical student-teacher mismatch: the teacher's inability to access long-term history prevents it from guiding the student on global temporal dependencies, effectively capping the student's context length. To resolve this, we propose Context Forcing, a novel framework that trains a long-context student via a long-context teacher. By ensuring the teacher is aware of the full generation history, we eliminate the supervision mismatch, enabling the robust training of models capable of long-term consistency. To make this computationally feasible for extreme durations (e.g., 2 minutes), we introduce a context management system that transforms the linearly growing context into a Slow-Fast Memory architecture, significantly reducing visual redundancy. Extensive results demonstrate that our method enables effective context lengths exceeding 20 seconds -- 2 to 10 times longer than state-of-the-art methods like LongLive and Infinite-RoPE. By leveraging this extended context, Context Forcing preserves superior consistency across long durations, surpassing state-of-the-art baselines on various long video evaluation metrics.

北京市昌平区探索星信息技术及软件开发工作室

京ICP备2026059466号
Context Forcing: Consistent Autoregressive Video Generation with Long Context | TensorX