{"ID":2838909,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.16047","arxiv_id":"2511.16047","title":"AMS-KV: Adaptive KV Caching in Multi-Scale Visual Autoregressive Transformers","abstract":"Visual autoregressive modeling (VAR) via next-scale prediction has emerged as a scalable image generation paradigm. While Key and Value (KV) caching in large language models (LLMs) has been extensively studied, next-scale prediction presents unique challenges, and KV caching design for next-scale based VAR transformers remains largely unexplored. A major bottleneck is the excessive KV memory growth with the increasing number of scales-severely limiting scalability. Our systematic investigation reveals that: (1) Attending to tokens from local scales significantly contributes to generation quality (2) Allocating a small amount of memory for the coarsest scales, termed as condensed scales, stabilizes multi-scale image generation (3) Strong KV similarity across finer scales is predominantly observed in cache-efficient layers, whereas cache-demanding layers exhibit weaker inter-scale similarity. Based on the observations, we introduce AMS-KV, a scale-adaptive KV caching policy for next-scale prediction in VAR models. AMS-KV prioritizes storing KVs from condensed and local scales, preserving the most relevant tokens to maintain generation quality. It further optimizes KV cache utilization and computational efficiency identifying cache-demanding layers through inter-scale similarity analysis. Compared to the vanilla next-scale prediction-based VAR models, AMS-KV reduces KV cache usage by up to 84.83% and self-attention latency by 60.48%. Moreover, when the baseline VAR-d30 model encounters out-of-memory failures at a batch size of 128, AMS-KV enables stable scaling to a batch size of 256 with improved throughput.","short_abstract":"Visual autoregressive modeling (VAR) via next-scale prediction has emerged as a scalable image generation paradigm. While Key and Value (KV) caching in large language models (LLMs) has been extensively studied, next-scale prediction presents unique challenges, and KV caching design for next-scale based VAR transformers...","url_abs":"https://arxiv.org/abs/2511.16047","url_pdf":"https://arxiv.org/pdf/2511.16047v1","authors":"[\"Boxun Xu\",\"Yu Wang\",\"Zihu Wang\",\"Peng Li\"]","published":"2025-11-20T05:10:12Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Transformer\",\"Large Language Model\",\"Language Model\"]","has_code":false}