AdaTP: Attention-Debiased Token Pruning for Video Large Language Models

• URL: http://arxiv.org/abs/2505.20100v1
• Date: Mon, 26 May 2025 15:08:37 GMT
• Title: AdaTP: Attention-Debiased Token Pruning for Video Large Language Models
• Authors: Fengyuan Sun, Leqi Shen, Hui Chen, Sicheng Zhao, Jungong Han, Guiguang Ding,
• Abstract summary: Video Large Language Models (Video LLMs) have achieved remarkable results in video understanding tasks.<n>They often suffer from heavy computational overhead due to the large number of visual tokens generated from multiple video frames.<n>AdaTP integrates two dedicated debiasing modules into the pipeline, targeting global attention bias and local attention bias.
• Score: 73.70241811066907
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Video Large Language Models (Video LLMs) have achieved remarkable results in video understanding tasks. However, they often suffer from heavy computational overhead due to the large number of visual tokens generated from multiple video frames. Existing visual token compression methods often rely on attention scores from language models as guidance. However, these scores exhibit inherent biases: global bias reflects a tendency to focus on the two ends of the visual token sequence, while local bias leads to an over-concentration on the same spatial positions across different frames. To address the issue of attention bias, we propose $\textbf{A}$ttention-$\textbf{D}$ebi$\textbf{a}$sed $\textbf{T}$oken $\textbf{P}$runing for Video Large Language Models ($\textbf{AdaTP}$), a novel token pruning pipeline for Video LLMs. AdaTP integrates two dedicated debiasing modules into the pipeline, targeting global attention bias and local attention bias, respectively. Without the need for additional training, our method significantly reduces the computational overhead of Video LLMs while retaining the performance of vanilla models. Extensive evaluation shows that AdaTP achieves state-of-the-art performance in various commonly used video understanding benchmarks. In particular, on LLaVA-OneVision-7B, AdaTP maintains performance without degradation while using only up to $27.3\%$ FLOPs compared to the vanilla model. Our code will be released soon.

Related papers

• PEVLM: Parallel Encoding for Vision-Language Models [4.777805570120456]
  We introduce textbfPEVLM, a fine-tuning-free parallel encoding method designed to enhance the prefilling efficiency of Vision-Language Models.<n>PEVLM partitions the input video into context blocks with a shared sink block, while preserving sequential position embeddings to align the attention weight distribution with that of Full-Attention.<n>Experiments demonstrate that PEVLM consistently outperforms existing parallel encoding approaches, achieving up to textbf7.47x speedup in attention computation and reducing end-to-end latency by textbf40%.
  arXiv  Detail & Related papers  (2025-06-24T14:14:52Z)
• REEF: Relevance-Aware and Efficient LLM Adapter for Video Understanding [2.309018557701645]
  Recent methods often compress memory banks to handle untemporal videos for video-level understanding.<n>To this, we designed video to compress un videos on a large scale using visual tokens.
  arXiv  Detail & Related papers  (2025-04-07T20:36:34Z)
• FastVID: Dynamic Density Pruning for Fast Video Large Language Models [38.267065642416554]
  We propose Density Pruning for Fast Video LLMs termed FastVID.<n>FastVID partitions videos into temporally ordered segments to preserve temporal structure.<n>Our method significantly reduces computational overhead while maintaining temporal and visual integrity.
  arXiv  Detail & Related papers  (2025-03-14T08:33:08Z)
• QuoTA: Query-oriented Token Assignment via CoT Query Decouple for Long Video Comprehension [86.0749609778104]
  We propose QuoTA, an ante-hoc training-free modular that extends existing large video-language models.<n>QuoTA strategically allocates frame-level importance scores based on query relevance.<n>We decouple the query through Chain-of-Thoughts reasoning to facilitate more precise LVLM-based frame importance scoring.
  arXiv  Detail & Related papers  (2025-03-11T17:59:57Z)
• Rethinking Homogeneity of Vision and Text Tokens in Large Vision-and-Language Models [29.611769371733672]
  We propose De Attention (D-Attn), a novel method that processes visual and textual embeddings differently.<n>D-Attn diagonalizes visual-to-visual self-attention, reducing computation from $mathcalO(|V|2)$ to $mathcalO(|V|)$ for $|V|$ visual embeddings without compromising performance.
  arXiv  Detail & Related papers  (2025-02-04T00:46:11Z)
• ST$^3$: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimming [14.937905258757635]
  $textbfST3$ is a framework designed to accelerate MLLM inference without retraining.<n>$textbfST3$ can be seamlessly integrated into existing pre-trained MLLMs.
  arXiv  Detail & Related papers  (2024-12-28T10:17:29Z)
• Looking Beyond Text: Reducing Language bias in Large Vision-Language Models via Multimodal Dual-Attention and Soft-Image Guidance [67.26434607115392]
  Large vision-language models (LVLMs) have achieved impressive results in various vision-language tasks. LVLMs suffer from hallucinations caused by language bias, leading to diminished focus on images and ineffective visual comprehension. We propose LACING to address the language bias of LVLMs with muLtimodal duAl-attention meChanIsm (MDA) aNd soft-image Guidance (IFG)
  arXiv  Detail & Related papers  (2024-11-21T16:33:30Z)
• Inference Optimal VLMs Need Fewer Visual Tokens and More Parameters [54.01228554126122]
  Vision Language Models (VLMs) have demonstrated strong capabilities across various visual understanding and reasoning tasks.<n>To reduce inference costs, one can either downsize the Large Language Models (LLMs) or reduce the number of input tokens needed to represent the image.<n>We take the first steps toward designing token compression algorithms tailored for high-compression settings.
  arXiv  Detail & Related papers  (2024-11-05T18:54:21Z)
• Attention Speaks Volumes: Localizing and Mitigating Bias in Language Models [15.53216696218776]
  We explore the internal mechanisms of how bias emerges in large language models (LLMs) when provided with ambiguous comparative prompts. We propose $textttATLAS$, a technique to localize bias to specific layers of the LLM by analyzing attention scores and then reduce bias by scaling attention in these biased layers.
  arXiv  Detail & Related papers  (2024-10-29T20:15:56Z)
• 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)
• Mitigating Representation Bias in Action Recognition: Algorithms and Benchmarks [76.35271072704384]
  Deep learning models perform poorly when applied to videos with rare scenes or objects. We tackle this problem from two different angles: algorithm and dataset. We show that the debiased representation can generalize better when transferred to other datasets and tasks.
  arXiv  Detail & Related papers  (2022-09-20T00:30:35Z)
• Video Understanding as Machine Translation [53.59298393079866]
  We tackle a wide variety of downstream video understanding tasks by means of a single unified framework. We report performance gains over the state-of-the-art on several downstream tasks including video classification (EPIC-Kitchens), question answering (TVQA), captioning (TVC, YouCook2, and MSR-VTT)
  arXiv  Detail & Related papers  (2020-06-12T14:07:04Z)

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.