Towards Memory-Efficient Autoregressive Video Generation via Instance-Specific Parametric Absorption

• URL: http://arxiv.org/abs/2607.00712v1
• Date: Wed, 01 Jul 2026 09:59:28 GMT
• Title: Towards Memory-Efficient Autoregressive Video Generation via Instance-Specific Parametric Absorption
• Abstract summary: Instance-Specific Parametric Absorption (ISPA) is a novel framework that shifts the KV cache compression from discarding to distilling.<n>We show ISPA can remove up to 50% of the KV cache with near-lossless visual quality.
• Score: 37.96485836820005
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Autoregressive (AR) streaming models have emerged as a powerful paradigm for long video generation. However, the linearly growing Key-Value (KV) cache poses a significant bottleneck, leading to memory overload and degraded inference throughput. A common compression method is to drop redundant KV tokens, which often breaks long-range dependencies, resulting in temporal flickering and identity loss. In this paper, we propose Instance-Specific Parametric Absorption (ISPA), a novel framework that shifts the KV cache compression from discarding to distilling. The core idea is to transit a subset of layers from Full-Attention (F-Layers) to memory-efficient Local-Attention (L-Layers) by "absorbing" historical context into the model's weights. Specifically, during a brief warmup phase, ISPA monitors the output discrepancy between global and local attention. At the transition point, we solve a closed-form least-squares problem to compute an instance-specific weight modulation that compensates for the missing history. Experiments across architectures (1.3B to 14B) demonstrate that ISPA can remove up to 50\% of the KV cache with near-lossless visual quality. We hope this perspective encourages future work to explore parametric memory consolidation beyond external token-level cache management for streaming generative models.

Related papers

• Attend Locally, Remember Linearly: Linear Attention as Cross-Frame Memory for Autoregressive Video Diffusion [61.57938553036056]
  ARL2 is a hybrid attention module that replaces quadratic cross-frame attention with a fixed-size recurrent state.<n>We decompose self-attention into two branches: an intra-frame softmax branch for spatial detail and local dependencies, and an inter-frame recurrent linear branch that maintains a fixed-size state for streaming context.<n>With 75% of layers replaced by hybrid linear attention, the model achieves up to 2.26 wall-clock speedup and 54% memory reduction, while maintaining comparable quality with improving temporal consistency.
  arXiv  Detail & Related papers  (2026-05-15T19:33:45Z)
• Latent-Condensed Transformer for Efficient Long Context Modeling [60.72493959155964]
  We propose Latent-Condensed Attention (LCA) for large language models.<n>LCA condenses context within MLA's latent space, where representation is disentangled into semantic latent vectors and positional keys.<n>LCA achieves up to 2.5$times$ prefilling speedup and 90% KV cache reduction at 128K context.
  arXiv  Detail & Related papers  (2026-04-14T08:40:31Z)
• Stochastic KV Routing: Enabling Adaptive Depth-Wise Cache Sharing [29.913403615975174]
  Serving transformer language models with high throughput requires caching Key-Values ( KVs) to avoid redundant computation.<n>The memory footprint of KV caching is significant and heavily impacts serving costs.<n>We propose a simple training approach: random cross-layer attention. During training, layers randomly choose to attend either to their own KV states or those of a preceding layer.
  arXiv  Detail & Related papers  (2026-04-03T14:56:17Z)
• Stateful KV Cache Management for LLMs: Balancing Space, Time, Accuracy, and Positional Fidelity [0.0]
  Key-Value (KV) cache is integral to efficient autoregressive inference in large language models (LLMs)<n>This paper examines the interplay between KV cache management strategies and the architectural context limits of models like meta-llama/Meta-Llama-3-8b-instruct.
  arXiv  Detail & Related papers  (2025-10-23T18:22:00Z)
• KV-Efficient VLA: A Method of Speed up Vision Language Model with RNN-Gated Chunked KV Cache [0.9238700679836854]
  Vision-Language-Action (VLA) models promise unified robotic perception and control, yet their scalability is constrained by the quadratic cost of attention and the unbounded growth of key-value (KV) memory during long-horizon inference.<n>We present KV-Efficient VLA, a model-agnostic memory compression framework that addresses these limitations by introducing a lightweight, training-friendly mechanism to selectively retain high-utility context.<n>Our method integrates seamlessly into existing autoregressive and hybrid VLA stacks, enabling scalable inference without modifying training pipelines or downstream control logic.
  arXiv  Detail & Related papers  (2025-09-20T02:04:24Z)
• LoRC: Low-Rank Compression for LLMs KV Cache with a Progressive Compression Strategy [59.1298692559785]
  Key-Value ( KV) cache is crucial component in serving transformer-based autoregressive large language models (LLMs) Existing approaches to mitigate this issue include: (1) efficient attention variants integrated in upcycling stages; (2) KV cache compression at test time; and (3) KV cache compression at test time. We propose a low-rank approximation of KV weight matrices, allowing plug-in integration with existing transformer-based LLMs without model retraining. Our method is designed to function without model tuning in upcycling stages or task-specific profiling in test stages.
  arXiv  Detail & Related papers  (2024-10-04T03:10:53Z)
• ThinK: Thinner Key Cache by Query-Driven Pruning [63.13363917871414]
  Large Language Models (LLMs) have revolutionized the field of natural language processing, achieving unprecedented performance across a variety of applications.<n>This paper focuses on the long-context scenario, addressing the inefficiencies in KV cache memory consumption during inference.<n>We propose ThinK, a novel query-dependent KV cache pruning method designed to minimize attention weight loss while selectively pruning the least significant channels.
  arXiv  Detail & Related papers  (2024-07-30T17:59:08Z)
• Efficient Inference of Vision Instruction-Following Models with Elastic Cache [76.44955111634545]
  We introduce Elastic Cache, a novel strategy for efficient deployment of instruction-following large vision-language models. We propose an importance-driven cache merging strategy to prune redundancy caches. For instruction encoding, we utilize the frequency to evaluate the importance of caches. Results on a range of LVLMs demonstrate that Elastic Cache not only boosts efficiency but also notably outperforms existing pruning methods in language generation.
  arXiv  Detail & Related papers  (2024-07-25T15:29:05Z)

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.