DyMU: Dynamic Merging and Virtual Unmerging for Efficient VLMs

• URL: http://arxiv.org/abs/2504.17040v2
• Date: Sat, 10 May 2025 22:42:28 GMT
• Title: DyMU: Dynamic Merging and Virtual Unmerging for Efficient VLMs
• Authors: Zhenhailong Wang, Senthil Purushwalkam, Caiming Xiong, Silvio Savarese, Heng Ji, Ran Xu,
• Abstract summary: We present DyMU, an efficient, training-free framework that reduces the computational burden of vision-language models (VLMs)<n>Our approach comprises two key components. First, Dynamic Token Merging (DToMe) reduces the number of visual token embeddings by merging similar tokens based on image complexity.<n>Second, Virtual Token Unmerging (VTU) simulates the expected token sequence for large language models (LLMs) by efficiently reconstructing the attention dynamics of a full sequence.
• Score: 124.52164183968145
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We present DyMU, an efficient, training-free framework that dynamically reduces the computational burden of vision-language models (VLMs) while maintaining high task performance. Our approach comprises two key components. First, Dynamic Token Merging (DToMe) reduces the number of visual token embeddings by merging similar tokens based on image complexity, addressing the inherent inefficiency of fixed-length outputs in vision transformers. Second, Virtual Token Unmerging (VTU) simulates the expected token sequence for large language models (LLMs) by efficiently reconstructing the attention dynamics of a full sequence, thus preserving the downstream performance without additional fine-tuning. Unlike previous approaches, our method dynamically adapts token compression to the content of the image and operates completely training-free, making it readily applicable to most state-of-the-art VLM architectures. Extensive experiments on image and video understanding tasks demonstrate that DyMU can reduce the average visual token count by 32%-85% while achieving comparable performance to full-length models across diverse VLM architectures, including the recently popularized AnyRes-based visual encoders. Furthermore, through qualitative analyses, we demonstrate that DToMe effectively adapts token reduction based on image complexity and, unlike existing systems, provides users more control over computational costs. Project page: https://mikewangwzhl.github.io/dymu/.

Related papers

• Co-Training Vision Language Models for Remote Sensing Multi-task Learning [68.15604397741753]
  Vision language models (VLMs) have achieved promising results in RS image understanding, grounding, and ultra-high-resolution (UHR) image reasoning.<n>We present RSCoVLM, a simple yet flexible VLM baseline for RS MTL.<n>We propose a unified dynamic-resolution strategy to address the diverse image scales inherent in RS imagery.
  arXiv  Detail & Related papers  (2025-11-26T10:55:07Z)
• FLoC: Facility Location-Based Efficient Visual Token Compression for Long Video Understanding [55.700832127331324]
  FLoC is an efficient visual token compression framework based on the facility location function.<n>Our method achieves remarkable efficiency gains by swiftly selecting a compact subset of tokens.<n>Our approach is training-free, model-agnostic, and query-agnostic, providing a versatile solution.
  arXiv  Detail & Related papers  (2025-10-31T17:29:39Z)
• ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution [71.69364653858447]
  Existing Multimodal Large Language Models (MLLMs) suffer from increased inference costs due to the additional vision tokens introduced by image inputs.<n>We propose Visual Consistency Learning (ViCO), a novel training algorithm that enables the model to represent images of varying complexities using different numbers of vision tokens.<n> Experimental results demonstrate that our method can reduce the number of vision tokens by up to 50% while maintaining the model's perception, reasoning, and OCR capabilities.
  arXiv  Detail & Related papers  (2025-10-14T17:58:10Z)
• Variation-aware Vision Token Dropping for Faster Large Vision-Language Models [24.952668143243542]
  Large vision-language models (LVLMs) have demonstrated remarkable capabilities in multimodal understanding tasks.<n> Token compression offers a direct solution by reducing the number of tokens to be processed, thereby improving computational efficiency.<n>We propose Variation-aware Vision Token Dropping (textiti.e., textbfV$2$Drop), which progressively removes visual tokens with minimal variation during LVLM inference.
  arXiv  Detail & Related papers  (2025-09-01T15:28:44Z)
• CrossLMM: Decoupling Long Video Sequences from LMMs via Dual Cross-Attention Mechanisms [16.41418610688371]
  We introduce CrossLMM, which substantially reduces visual token quantity with minimal performance degradation.<n>We also introduce a text-to-visual cross-attention mechanism, for which the text tokens are enhanced through interaction with the original visual tokens.<n>Our approach achieves comparable or superior performance across diverse video-based Large Language Models benchmarks.
  arXiv  Detail & Related papers  (2025-05-22T17:59:53Z)
• InternVL-X: Advancing and Accelerating InternVL Series with Efficient Visual Token Compression [1.8893427856534721]
  We propose InternVL-X, which outperforms the InternVL model in both performance and efficiency.<n>By utilizing 20% or fewer visual tokens, InternVL-X achieves state-of-the-art performance on 7 public MLLM benchmarks, and improves the average metric by 2.34% across 12 tasks.
  arXiv  Detail & Related papers  (2025-03-27T09:31:35Z)
• Skip-Vision: Efficient and Scalable Acceleration of Vision-Language Models via Adaptive Token Skipping [13.846838416902575]
  A key bottleneck stems from the proliferation of visual tokens required for fine-grained image understanding.<n>We propose Skip-Vision, a unified framework addressing both training and inference inefficiencies in vision-language models.<n> Experimental results demonstrate that Skip-Vision reduces training time by up to 35%, inference FLOPs by 75%, and latency by 45%.
  arXiv  Detail & Related papers  (2025-03-26T04:16:48Z)
• Token-Efficient Long Video Understanding for Multimodal LLMs [101.70681093383365]
  STORM is a novel architecture incorporating a dedicated temporal encoder between the image encoder and the Video-LLMs.<n>We show that STORM achieves state-of-the-art results across various long video understanding benchmarks.
  arXiv  Detail & Related papers  (2025-03-06T06:17:38Z)
• ContextFormer: Redefining Efficiency in Semantic Segmentation [48.81126061219231]
  Convolutional methods, although capturing local dependencies well, struggle with long-range relationships.<n>Vision Transformers (ViTs) excel in global context capture but are hindered by high computational demands.<n>We propose ContextFormer, a hybrid framework leveraging the strengths of CNNs and ViTs in the bottleneck to balance efficiency, accuracy, and robustness for real-time semantic segmentation.
  arXiv  Detail & Related papers  (2025-01-31T16:11:04Z)
• Dynamic-LLaVA: Efficient Multimodal Large Language Models via Dynamic Vision-language Context Sparsification [27.84464349480918]
  We propose Dynamic vision-language context sparsification framework Dynamic-LLaVA. It dynamically reduces the redundancy of vision context in the prefill stage. It decreases the memory and computation overhead of the generated language context during decoding.
  arXiv  Detail & Related papers  (2024-12-01T16:32:31Z)
• ADEM-VL: Adaptive and Embedded Fusion for Efficient Vision-Language Tuning [38.26304604660713]
  ADEM-VL is an efficient vision-language method that tunes models based on pretrained large language models. Our framework surpasses existing methods by an average accuracy of 0.77% on ScienceQA dataset.
  arXiv  Detail & Related papers  (2024-10-23T11:31:06Z)
• EMMA: Efficient Visual Alignment in Multi-Modal LLMs [56.03417732498859]
  EMMA is a lightweight cross-modality module designed to efficiently fuse visual and textual encodings. EMMA boosts performance across multiple tasks by up to 9.3% while significantly improving robustness against hallucinations.
  arXiv  Detail & Related papers  (2024-10-02T23:00:31Z)
• VideoLLM-MoD: Efficient Video-Language Streaming with Mixture-of-Depths Vision Computation [66.00245701441547]
  We introduce a novel approach to reduce vision compute by leveraging redundant vision tokens "skipping layers" rather than decreasing the number of vision tokens. Our method, VideoLLM-MoD, is inspired by mixture-of-depths LLMs and addresses the challenge of numerous vision tokens in long-term or streaming video.
  arXiv  Detail & Related papers  (2024-08-29T17:21:58Z)
• Towards Semantic Equivalence of Tokenization in Multimodal LLM [149.11720372278273]
  Vision tokenization is essential for semantic alignment between vision and language.<n>This paper proposes a novel dynamic Semantic-Equivalent Vision Tokenizer (SeTok)<n>SeTok groups visual features into semantic units via a dynamic clustering algorithm.<n>The resulting vision tokens effectively preserve semantic integrity and capture both low-frequency and high-frequency visual features.
  arXiv  Detail & Related papers  (2024-06-07T17:55:43Z)
• Efficient Visual State Space Model for Image Deblurring [83.57239834238035]
  Convolutional neural networks (CNNs) and Vision Transformers (ViTs) have achieved excellent performance in image restoration. We propose a simple yet effective visual state space model (EVSSM) for image deblurring.
  arXiv  Detail & Related papers  (2024-05-23T09:13:36Z)

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.