# Q Cache: Visual Attention is Valuable in Less than Half of Decode Layers for Multimodal Large Language Model

arXiv: https://arxiv.org/abs/2602.01901 · HTML: https://mlsanity.com/paper/2602.01901/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 28 of its 44 references that are in the map.
- First posted on arXiv 2 Feb 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Visual token pruning and efficiency](https://mlsanity.com/mllm/visual-token-pruning-and-efficiency/)
- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

## Builds on (most cited first, 28 of 28)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://mlsanity.com/paper/2311.16502/) (2023)
- [Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering](https://mlsanity.com/paper/2209.09513/) (2022)
- [Towards VQA Models That Can Read](https://mlsanity.com/paper/1904.08920/) (2019)
- [Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding](https://mlsanity.com/paper/2306.02858/) (2023)
- [Microsoft COCO Captions: Data Collection and Evaluation Server](https://mlsanity.com/paper/1504.00325/) (2015)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models](https://mlsanity.com/paper/1505.04870/) (2015)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [A Diagram Is Worth A Dozen Images](https://mlsanity.com/paper/1603.07396/) (2016)
- [SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inference](https://mlsanity.com/paper/2410.04417/) (2024)
- [InternLM-XComposer2: Mastering Free-form Text-Image Composition and Comprehension in Vision-Language Large Model](https://mlsanity.com/paper/2401.16420/) (2024)
- [PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction](https://mlsanity.com/paper/2410.17247/) (2024)
- [Honeybee: Locality-enhanced Projector for Multimodal LLM](https://mlsanity.com/paper/2312.06742/) (2023)
- [TokenPacker: Efficient Visual Projector for Multimodal LLM](https://mlsanity.com/paper/2407.02392/) (2024)
- [Boosting Multimodal Large Language Models with Visual Tokens Withdrawal for Rapid Inference](https://mlsanity.com/paper/2405.05803/) (2024)
- [CrossGET: Cross-Guided Ensemble of Tokens for Accelerating Vision-Language Transformers](https://mlsanity.com/paper/2305.17455/) (2023)
- [VL-Cache: Sparsity and Modality-Aware KV Cache Compression for Vision-Language Model Inference Acceleration](https://mlsanity.com/paper/2410.23317/) (2024)
- [ST3: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimming](https://mlsanity.com/paper/2412.20105/) (2024)
- [Multi-task Visual Grounding with Coarse-to-Fine Consistency Constraints](https://mlsanity.com/paper/2501.06710/) (2025)
- [FALIP: Visual Prompt as Foveal Attention Boosts CLIP Zero-Shot Performance](https://mlsanity.com/paper/2407.05578/) (2024)

## Abstract (arXiv)

Multimodal large language models (MLLMs) are plagued by exorbitant inference costs attributable to the profusion of visual tokens within the vision encoder. The redundant visual tokens engenders a substantial computational load and key-value (KV) cache footprint bottleneck. Existing approaches focus on token-wise optimization, leveraging diverse intricate token pruning techniques to eliminate non-crucial visual tokens. Nevertheless, these methods often unavoidably undermine the integrity of the KV cache, resulting in failures in long-text generation tasks. To this end, we conduct an in-depth investigation towards the attention mechanism of the model from a new perspective, and discern that attention within more than half of all decode layers are semantic similar. Upon this finding, we contend that the attention in certain layers can be streamlined by inheriting the attention from their preceding layers. Consequently, we propose Lazy Attention, an efficient attention mechanism that enables cross-layer sharing of similar attention patterns. It ingeniously reduces layer-wise redundant computation in attention. In Lazy Attention, we develop a novel layer-shared cache, Q Cache, tailored for MLLMs, which facilitates the reuse of queries across adjacent layers. In particular, Q Cache is lightweight and fully compatible with existing inference frameworks, including Flash Attention and KV cache. Additionally, our method is highly flexible as it is orthogonal to existing token-wise techniques and can be deployed independently or combined with token pruning approaches. Empirical evaluations on multiple benchmarks demonstrate that our method can reduce KV cache usage by over 35% and achieve 1.5x throughput improvement, while sacrificing only approximately 1% of performance on various MLLMs. Compared with SOTA token-wise methods, our technique achieves superior accuracy preservation.