Pyramid Forcing: Head-Aware Pyramid KV Cache Policy for High-Quality Long Video Generation

• URL: http://arxiv.org/abs/2605.13111v1
• Date: Wed, 13 May 2026 07:23:02 GMT
• Title: Pyramid Forcing: Head-Aware Pyramid KV Cache Policy for High-Quality Long Video Generation
• Abstract summary: Pyramid Forcing is a head-aware KVCache framework that identifies head types offline and assigns behavior-specific cache policies.<n> Experiments show that Pyramid Forcing consistently improves long-horizon generation quality on VBench-Long.
• Score: 14.413680197991356
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Autoregressive video generation enables streaming and open-ended long video synthesis, but still suffers from long-term degradation caused by accumulated errors. Existing KVCache strategies usually apply unified historical-frame retention, implicitly assuming homogeneous historical dependencies across attention heads. We revisit historical-frame attention and reveal three distinct head types: Anchor Heads require broad long-range context, Wave Heads exhibit periodic temporal dependencies, and Veil Heads focus on initial and adjacent frames. Based on this finding, we propose Pyramid Forcing, a head-aware pyramidal KVCache framework that identifies head types offline, assigns behavior-specific cache policies, and supports heterogeneous cache lengths via efficient ragged-cache attention. Experiments on Self Forcing and Causal Forcing show that Pyramid Forcing consistently improves long-horizon generation quality on VBench-Long, increasing the 60-second Self Forcing score from 77.87 to 81.21 while enhancing motion dynamics, visual fidelity, and semantic consistency. Project: https://if-lab-pku.github.io/Pyramid-Forcing/.

Related papers

• TetherCache: Stabilizing Autoregressive Long-Form Video Generation with Gated Recall and Trusted Alignment [51.33418612284208]
  We propose TetherCache, a training-free and plug-and-play cache management strategy for drift-resistant long video generation.<n>Gated Recall with Attention-Diversity Balancing selects long-range memory frames using a gated score.<n>TAME lightly edits newly recalled memory tokens by aligning their statistics to a trusted context distribution.
  arXiv  Detail & Related papers  (2026-06-11T08:16:08Z)
• FadeMem: Distance-Aware Memory Consolidation for Autoregressive Video Diffusion [59.207505503284715]
  FadeMem organizes historical KV blocks into a temporal hierarchy under a fixed cache budget.<n>New history is inserted as fine-grained entries, while older adjacent entries are progressively merged.<n>Experiments show improved subject consistency, background stability, and temporal coherence over existing bounded-cache strategies.
  arXiv  Detail & Related papers  (2026-06-09T10:22:18Z)
• DySink: Dynamic Frame Sinks for Autoregressive Long Video Generation [51.58559206569209]
  We propose DySink, a retrieval-based framework that maintains a compact memory bank and selects visually relevant historical frames as dynamic frame sinks.<n> Experiments on minute-long videos show that DySink consistently improves dynamic degree over strong baselines while also achieving higher temporal quality.
  arXiv  Detail & Related papers  (2026-05-20T11:01:01Z)
• Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion [25.555611454522126]
  We present textbfFocused Forcing, a training-free KV selection method that focuses cached history along both generated-frame and head dimensions.<n>For each generated frame, Focused Forcing preserves the most relevant and distinctive historical frames.<n>Across multiple autoregressive generation paradigms, Focused Forcing achieves up to $textbf1.48times$ end-to-end acceleration without training.
  arXiv  Detail & Related papers  (2026-05-18T12:58:13Z)
• Echo-Forcing: A Scene Memory Framework for Interactive Long Video Generation [48.476317015122625]
  Echo-Forcing is a training-free scene memory framework for interactive long video generation.<n>It supports smooth transitions, hard cuts, and long-range scene recall under a bounded cache budget.
  arXiv  Detail & Related papers  (2026-05-15T14:33:09Z)
• Head Forcing: Long Autoregressive Video Generation via Head Heterogeneity [21.347969441374506]
  Autoregressive video diffusion models support real-time synthesis but suffer from error accumulation and context loss over long horizons.<n>We propose Head Forcing, a training-free framework that assigns each head type a tailored KV cache strategy.<n>Without additional training, Head Forcing extends generation from 5 seconds to minute-level duration, supports multi-prompt interactive synthesis, and consistently outperforms existing baselines.
  arXiv  Detail & Related papers  (2026-05-14T07:27:39Z)
• LongStream: Long-Sequence Streaming Autoregressive Visual Geometry [18.46781332515933]
  LongStream is a novel gauge-decoupled streaming visual geometry model for metric-scale scene reconstruction.<n>It reformulates long-range extrapolation into a constant-difficulty local task.<n>It delivers stable, metric-scale reconstruction over kilometer-scale sequences at 18 FPS.
  arXiv  Detail & Related papers  (2026-02-13T18:30:51Z)
• Flow caching for autoregressive video generation [72.10021661412364]
  We present FlowCache, the first caching framework specifically designed for autoregressive video generation.<n>Our method achieves remarkable speedups of 2.38 times on MAGI-1 and 6.7 times on SkyReels-V2, with negligible quality degradation.
  arXiv  Detail & Related papers  (2026-02-11T13:11:04Z)
• Fast Autoregressive Video Diffusion and World Models with Temporal Cache Compression and Sparse Attention [37.91838955436801]
  Autoregressive video diffusion models enable streaming generation, opening the door to long-form synthesis, video world models, and interactive neural game engines.<n>As generation progresses, the KV cache grows, causing both increasing latency and escalating GPU memory, which in turn restricts usable temporal context and harms long-range consistency.<n>We propose a unified, training-free attention framework for autoregressive diffusion: TempCache compresses the KV cache via temporal correspondence to bound cache growth; AnnCA accelerates cross-attention by selecting frame-relevant prompt tokens using fast approximate nearest neighbor matching; and AnnSA sparsifies self-attention by restricting each query
  arXiv  Detail & Related papers  (2026-02-02T08:31:21Z)
• Efficient Autoregressive Video Diffusion with Dummy Head [35.72735205078644]
  We propose Dummy Forcing, a method to control context accessibility across different heads.<n>Specifically, the proposed heterogeneous memory allocation reduces head-wise context redundancy, accompanied by dynamic head programming.<n>Without additional training, our Dummy Forcing delivers up to 2.0x speedup over the baseline, supporting video generation at 24.3 FPS with less than 0.5% quality drop.
  arXiv  Detail & Related papers  (2026-01-28T11:20:43Z)
• LoL: Longer than Longer, Scaling Video Generation to Hour [50.945885467651216]
  This work achieves the first demonstration of real-time, streaming, and infinite-length video generation with little quality decay.<n>As an illustration, we generate continuous videos up to 12 hours in length, which, to our knowledge, is among the longest publicly demonstrated results in streaming video generation.
  arXiv  Detail & Related papers  (2026-01-23T17:21:35Z)
• HeteroCache: A Dynamic Retrieval Approach to Heterogeneous KV Cache Compression for Long-Context LLM Inference [14.17979669446161]
  We propose HeteroCache, a training-free dynamic compression framework.<n>We show that HeteroCache achieves state-of-the-art performance on multiple long-context benchmarks and accelerates decoding by up to $3times$ compared to the original model in the 224K context.
  arXiv  Detail & Related papers  (2026-01-20T07:35:06Z)
• PackCache: A Training-Free Acceleration Method for Unified Autoregressive Video Generation via Compact KV-Cache [61.57938553036056]
  We introduce PackCache, a training-free KV-cache management method that compacts the KV cache through three coordinated mechanisms.<n>In terms of efficiency, PackCache accelerates end-to-end generation by 1.7-2.2x on 48-frame long sequences.
  arXiv  Detail & Related papers  (2026-01-07T19:51:06Z)
• Lookahead Anchoring: Preserving Character Identity in Audio-Driven Human Animation [75.71558917038838]
  Lookahead Anchoring prevents identity drift during temporal autoregressive generation.<n>It transforms from fixed boundaries into directional beacons.<n>It also enables self-key-framing, where the reference image serves as the lookahead target.
  arXiv  Detail & Related papers  (2025-10-27T17:50:19Z)
• LongLive: Real-time Interactive Long Video Generation [68.45945318075432]
  LongLive is a frame-level autoregressive framework for real-time and interactive long video generation.<n>LongLive sustains 20.7 FPS on a single NVIDIA H100, achieves strong performance on VBench in both short and long videos.
  arXiv  Detail & Related papers  (2025-09-26T17:48:24Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.