# HeatKV: Head-tuned KV-cache Compression for Visual Autoregressive Modeling

arXiv: https://arxiv.org/abs/2605.14877 · HTML: https://mlsanity.com/paper/2605.14877/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Fast autoregressive visual generation, one of 12 directions in Efficient LLM inference and serving.
- Builds on 37 of its 40 references that are in the map.
- First posted on arXiv 14 May 2026.

## Placed in

- Efficient LLM inference and serving / [Fast autoregressive visual generation](https://mlsanity.com/inference/fast-autoregressive-visual-generation/)

## Builds on (most cited first, 30 of 37)

- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Diffusion Models Beat GANs on Image Synthesis](https://mlsanity.com/paper/2105.05233/) (2021)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [Neural Discrete Representation Learning](https://mlsanity.com/paper/1711.00937/) (2017)
- [Taming Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2012.09841/) (2020)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Scaling Autoregressive Models for Content-Rich Text-to-Image Generation](https://mlsanity.com/paper/2206.10789/) (2022)
- [Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction](https://mlsanity.com/paper/2404.02905/) (2024)
- [GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment](https://mlsanity.com/paper/2310.11513/) (2023)
- [Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis](https://mlsanity.com/paper/2306.09341/) (2023)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models](https://mlsanity.com/paper/2306.14048/) (2023)
- [Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation](https://mlsanity.com/paper/2406.06525/) (2024)
- [Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation](https://mlsanity.com/paper/2310.05737/) (2023)
- [StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets](https://mlsanity.com/paper/2202.00273/) (2022)
- [Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis](https://mlsanity.com/paper/2412.04431/) (2024)
- [Ada-KV: Optimizing KV Cache Eviction by Adaptive Budget Allocation for Efficient LLM Inference](https://mlsanity.com/paper/2407.11550/) (2024)
- [HART: Efficient Visual Generation with Hybrid Autoregressive Transformer](https://mlsanity.com/paper/2410.10812/) (2024)
- [Image and Video Tokenization with Binary Spherical Quantization](https://mlsanity.com/paper/2406.07548/) (2024)
- [FlowAR: Scale-wise Autoregressive Image Generation Meets Flow Matching](https://mlsanity.com/paper/2412.15205/) (2024)
- [Parallelized Autoregressive Visual Generation](https://mlsanity.com/paper/2412.15119/) (2024)
- [InfinityStar: Unified Spacetime AutoRegressive Modeling for Visual Generation](https://mlsanity.com/paper/2511.04675/) (2025)
- [Visual Autoregressive Modeling for Image Super-Resolution](https://mlsanity.com/paper/2501.18993/) (2025)
- [FastVAR: Linear Visual Autoregressive Modeling via Cached Token Pruning](https://mlsanity.com/paper/2503.23367/) (2025)
- [FlexVAR: Flexible Visual Autoregressive Modeling without Residual Prediction](https://mlsanity.com/paper/2502.20313/) (2025)

## Abstract (arXiv)

Visual Autoregressive (VAR) models have recently demonstrated impressive image generation quality while maintaining low latency. However, they suffer from severe KV-cache memory constraints, often requiring gigabytes of memory per generated image. We introduce HeatKV, a novel compression method that adapts cache allocation in each head based on its attention to previously generated scales. Using a small offline calibration set, the attention heads are ranked according to their attention scores over prior scales. Based on this ranking, we construct a static pruning schedule tailored to a given memory budget. Applied to the Infinity-2B model, HeatKV achieves 2 × higher compression ratio in memory allocation for KV cache compared to existing methods, while maintaining similar or better image fidelity, prompt alignment and human perception score. Our method achieves a new state-of-the-art (SOTA) for VAR model KV-cache compression, showcasing the effectiveness of fine-grained, head-specific cache allocation. Code and calibration script available at https://github.com/arm-research/heatkv.