LayerRecall：用於影片生成中長時程一致性的狀態條件化記憶路由器

摘要

自迴歸影片擴散透過從有限的近期上下文生成片段，實現可擴展的長影片生成。雖然基於近期性的快取能保持局部連續性，但當主體、物體、場景或屬性重新出現時，它會逐出所需的歷史線索。現有的記憶機制讓模型得以接觸非局部歷史，但僅有存取並不能確保有效利用。我們的分析揭示，影片DiT層對當前、近期與遙遠上下文表現出不同的偏好，這表明長程記憶需要同時決定檢索什麼以及在何處使用。我們提出LayerRecall，這是一個以當前狀態為條件、具層級選擇性的記憶路由器，它檢索相關的歷史K/V狀態，並僅將其注入主幹網路特有的記憶敏感層，同時在其他層保留局部注意力。為了減少對稀缺的高品質長視野影片及明確記憶配置標籤的依賴，我們進一步提出跨視野預測匹配（CHPM），該方法在預測空間中使用特權長上下文參考來監督有限記憶路由器。在100個多鏡頭評估提示中，LayerRecall在MemoBench和MovieBench上取得最佳整體結果，同時在VBench-Long上與其主幹網路表現相當，展現了在不犧牲局部連續性的前提下更強的長程恢復能力。定性分析進一步揭示記憶引導的自我修正機制，即最初不匹配的局部屬性會恢復至其歷史外觀，同時不會重置正在進行的運動或場景結構。額外分析顯示其具備跨主幹網路的可攜性以及可忽略的推論開銷。

English

Autoregressive video diffusion enables scalable long-video generation by producing chunks from a bounded recent context. While recency-based caching preserves local continuity, it evicts historical cues needed when subjects, objects, scenes, or attributes reappear. Existing memory mechanisms expose models to nonlocal history, but access alone does not ensure effective use. Our analysis reveals that video DiT layers exhibit distinct preferences for current, recent, and distant context, suggesting that long-range memory requires deciding both what to retrieve and where to use it. We introduce LayerRecall, a current-conditioned, layer-selective memory router that retrieves relevant historical K/V states and injects them only into backbone-specific memory-sensitive layers while preserving local attention elsewhere. To reduce reliance on scarce high-quality long-horizon videos and explicit memory-allocation labels, we further propose Cross-Horizon Prediction Matching (CHPM), which uses a privileged long-context reference to supervise the bounded-memory router in prediction space. Across 100 multi-shot evaluation prompts, LayerRecall achieves the best overall results on MemoBench and MovieBench while matching its backbone on VBench-Long, demonstrating stronger long-range recovery without sacrificing local continuity. Qualitative analyses further reveal memory-guided self-correction, whereby initially mismatched local attributes return to their historical appearance without resetting ongoing motion or scene structure. Additional analyses show cross-backbone portability and negligible inference overhead.