# Efficient Inference of Vision Instruction-Following Models with Elastic Cache

arXiv: https://arxiv.org/abs/2407.18121 · HTML: https://mlsanity.com/paper/2407.18121/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 27 of its 34 references that are in the map.
- First posted on arXiv 25 Jul 2024.

## Placed in

- Multimodal LLMs / vision-language models / [Visual token pruning and efficiency](https://mlsanity.com/mllm/visual-token-pruning-and-efficiency/)

## Builds on (most cited first, 27 of 27)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models](https://mlsanity.com/paper/2211.10438/) (2022)
- [MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://mlsanity.com/paper/2308.02490/) (2023)
- [HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face](https://mlsanity.com/paper/2303.17580/) (2023)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [A Simple and Effective Pruning Approach for Large Language Models](https://mlsanity.com/paper/2306.11695/) (2023)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [LLM-Pruner: On the Structural Pruning of Large Language Models](https://mlsanity.com/paper/2305.11627/) (2023)
- [FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU](https://mlsanity.com/paper/2303.06865/) (2023)
- [MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Action](https://mlsanity.com/paper/2303.11381/) (2023)
- [Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs](https://mlsanity.com/paper/2310.01801/) (2023)
- [Scissorhands: Exploiting the Persistence of Importance Hypothesis for LLM KV Cache Compression at Test Time](https://mlsanity.com/paper/2305.17118/) (2023)
- [DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scale](https://mlsanity.com/paper/2207.00032/) (2022)
- [Learning to Compress Prompts with Gist Tokens](https://mlsanity.com/paper/2304.08467/) (2023)
- [MultiModal-GPT: A Vision and Language Model for Dialogue with Humans](https://mlsanity.com/paper/2305.04790/) (2023)
- [Octopus: Embodied Vision-Language Programmer from Environmental Feedback](https://mlsanity.com/paper/2310.08588/) (2023)
- [Chain-of-Spot: Interactive Reasoning Improves Large Vision-Language Models](https://mlsanity.com/paper/2403.12966/) (2024)

## Built on by (most cited first, 24 of 24)

- [Oryx MLLM: On-Demand Spatial-Temporal Understanding at Arbitrary Resolution](https://mlsanity.com/paper/2409.12961/) (2024)
- [Ola: Pushing the Frontiers of Omni-Modal Language Model](https://mlsanity.com/paper/2502.04328/) (2025)
- [LiveVLM: Efficient Online Video Understanding via Streaming-Oriented KV Cache and Retrieval](https://mlsanity.com/paper/2505.15269/) (2025)
- [ZipVL: Efficient Large Vision-Language Models with Dynamic Token Sparsification](https://mlsanity.com/paper/2410.08584/) (2024)
- [SparseMM: Head Sparsity Emerges from Visual Concept Responses in MLLMs](https://mlsanity.com/paper/2506.05344/) (2025)
- [ModServe: Modality- and Stage-Aware Resource Disaggregation for Scalable Multimodal Model Serving](https://mlsanity.com/paper/2502.00937/) (2025)
- [Mixing Importance with Diversity: Joint Optimization for KV Cache Compression in Large Vision-Language Models](https://mlsanity.com/paper/2510.20707/) (2025)
- [PrefixKV: Adaptive Prefix KV Cache is What Vision Instruction-Following Models Need for Efficient Generation](https://mlsanity.com/paper/2412.03409/) (2024)
- [Efficient LLaMA-3.2-Vision by Trimming Cross-attended Visual Features](https://mlsanity.com/paper/2504.00557/) (2025)
- [Efficient Inference for Large Vision-Language Models: Bottlenecks, Techniques, and Prospects](https://mlsanity.com/paper/2604.05546/) (2026)
- [ElasticMM: Efficient Multimodal LLMs Serving with Elastic Multimodal Parallelism](https://mlsanity.com/paper/2507.10069/) (2025)
- [IKOD: Mitigating Visual Attention Degradation in Large Vision-Language Models](https://mlsanity.com/paper/2508.03469/) (2025)
- [Multi-Cue Adaptive Visual Token Pruning for Large Vision-Language Models](https://mlsanity.com/paper/2503.08019/) (2025)
- [PruneHal: Reducing Hallucinations in Multi-modal Large Language Models through Adaptive KV Cache Pruning](https://mlsanity.com/paper/2510.19183/) (2025)
- [Reroute, Don't Remove: Recoverable Visual Token Routing for Vision-Language Models](https://mlsanity.com/paper/2606.12412/) (2026)
- [Streamline Without Sacrifice -- Squeeze out Computation Redundancy in LMM](https://mlsanity.com/paper/2505.15816/) (2025)
- [Enabling Disaggregated Multi-Stage MLLM Inference via GPU-Internal Scheduling and Resource Sharing](https://mlsanity.com/paper/2512.17574/) (2025)
- [TABED: Test-Time Adaptive Ensemble Drafting for Robust Speculative Decoding in LVLMs](https://mlsanity.com/paper/2601.20357/) (2026)
- [Parallel In-context Learning for Large Vision Language Models](https://mlsanity.com/paper/2603.16092/) (2026)
- [SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning](https://mlsanity.com/paper/2603.23483/) (2026)
- [TCM-Serve: Modality-aware Scheduling for Multimodal Large Language Model Inference](https://mlsanity.com/paper/2603.26498/) (2026)
- [Make Your LVLM KV Cache More Lightweight](https://mlsanity.com/paper/2605.00789/) (2026)
- [TGV-KV: Text-Grounded KV Eviction for Vision-Language Models](https://mlsanity.com/paper/2606.03075/) (2026)
- [TwinPrune: Density-Aware Two-Phase Token Pruning for Vision-Language Models](https://mlsanity.com/paper/2610.17530/) (2026)

## Abstract (arXiv)

In the field of instruction-following large vision-language models (LVLMs), the efficient deployment of these models faces challenges, notably due to the high memory demands of their key-value (KV) caches. Conventional cache management strategies for LLMs focus on cache eviction, which often fails to address the specific needs of multimodal instruction-following models. Recognizing this gap, in this paper, we introduce Elastic Cache, a novel approach that benefits from applying distinct acceleration methods for instruction encoding and output generation stages. We investigate the metrics of importance in different stages and propose an importance-driven cache merging strategy to prune redundancy caches. Instead of discarding less important caches, our strategy identifies important key/value vectors as anchor points. Surrounding less important caches are then merged with these anchors, enhancing the preservation of contextual information in the KV caches while yielding an arbitrary acceleration ratio. For instruction encoding, we utilize the frequency to evaluate the importance of caches. Regarding output generation, we prioritize tokens based on their distance with an offset, by which both the initial and most recent tokens are retained. Results on a range of LVLMs demonstrate that Elastic Cache not only boosts efficiency but also notably outperforms existing pruning methods in language generation across various tasks. Code is available at https://github.com/liuzuyan/ElasticCache