Elamur Research⌕
☰
home / papers / 2026 / TrimTokenator: Towards Adaptive Visual Token Pruning for Large Multimodal Models

TrimTokenator: Towards Adaptive Visual Token Pruning for Large Multimodal Models

Paper · 2026 · Hao Zhang, Lyu Mengsi, Chenrui He et al.

Summary

Large multimodal models turn an image into thousands of visual tokens. Because self-attention cost grows quadratically with sequence length, these tokens make inference slow and memory-hungry. The authors first compare the two modalities through attention patterns, semantic distribution and repeated information. They find that visual tokens are far more redundant than textual ones, so they keep all text tokens and prune only the visual stream.

Their training-free method works in two steps:

  1. It scores each visual token by its average mutual information with the text tokens and discards the ones poorly aligned with the prompt.
  2. From the remaining tokens, a greedy procedure called Greedy RepMax keeps a subset that is spread out in embedding space. This approximates an NP-hard objective that maximizes expected pairwise distance.

On LLaVA-1.5-7B and LLaVA-NEXT-7B, the approach removes 88.9% of tokens with only small accuracy losses and speeds up inference by 56.7%. It beats attention-based, calibration-based and min-distance pruning baselines. The work matters for running LMMs on limited hardware or in latency-critical settings.

Key points

  • Baselines compared include ToMe, FastV, SparseVLM, PDrop, VisionZip and DivPrune. Benchmarks include ChartQA, COCO, MMB, MME, MMU and NoCaps.
  • On LLaVA-NEXT-7B keeping 320 of 2880 tokens, the method reaches an MME score of 1452.37. This compares with 1447.05 for DivPrune and 1519.30 for the unpruned model.
  • On LLaVA-NEXT-7B with 288 retained tokens on MME, the method reduces:
    • peak memory from 15.53 G to 13.63 G;
    • latency from 235.2 ms to 100.13 ms, with pruning overhead included.
  • With 64 retained tokens on LLaVA-1.5-7B, the method outperforms DivPrune* (DivPrune plus the paper's MI filter), scoring 61.34 vs 60.58 on MMB. Swapping the order of the two stages in the ablation causes a sharp drop (MME 1114.77).
  • The method was not evaluated on very large models (~70B parameters) because of compute limits.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🔬Beijing Academy of Artificial IntelligenceBeijing · China32 works · 7,511 citations

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksComputer Vision and Pattern Recognition · 2,869 works

🧭 Topics 2Research subjects and languages the work is about.Vision & language · 657 worksEfficient models & compression · 242 works

Cite

Cite as Hao Zhang, Lyu Mengsi, Chenrui He, Yulong Ao and Yonghua Lin. 2026. TrimTokenator: Towards Adaptive Visual Token Pruning for Large Multimodal Models. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1633

@inproceedings{zhang2026trimtokenator,
  title = {{TrimTokenator: Towards Adaptive Visual Token Pruning for Large Multimodal Models}},
  author = {Hao Zhang, Lyu Mengsi, Chenrui He, Yulong Ao and Yonghua Lin},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1633},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1633}
}

References

83 references, 23 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lada A Adamic and Bernardo A Huberman. 2002. Zipf’s law and the internet. Glottometrics, 3(1):143– 150. find
  2. Harsh Agrawal, Karan Desai, Yufei Wang, Xinlei Chen, Rishabh Jain, Mark Johnson, Dhruv Batra, Devi Parikh, Stefan Lee, and Peter Anderson. 2019. Nocaps: Novel object captioning at scale. In Proceedings of the IEEE/CVF international conference on computer vision, pages 8948–8957. find
  3. Saeed Ranjbar Alvar, Gursimran Singh, Mohammad Akbari, and Yong Zhang. 2025. Divprune: Diversity-based visual token pruning for large multimodal models. arXiv preprint arXiv:2503.02175. find
  4. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923. — Qwen2.5-VL Technical Report
  5. Christopher M Bishop and Nasser M Nasrabadi. 2006. Pattern recognition and machine learning, volume 4. Springer. find
  6. Yevgen Chebotar, Quan Vuong, Karol Hausman, Fei Xia, Yao Lu, Alex Irpan, Aviral Kumar, Tianhe Yu, Alexander Herzog, Karl Pertsch, and 1 others. 2023. Q-transformer: Scalable offline reinforcement learning via autoregressive q-functions. In Conference on Robot Learning, pages 3909–3928. PMLR. find
  7. Kevin Chen, Jiayuan Li, Yang Li, and Zhen Li. 2023. A survey on efficient transformer models. arXiv preprint. find
  8. Liang Chen, Haozhe Zhao, Tianyu Liu, Shuai Bai, Junyang Lin, Chang Zhou, and Baobao Chang. 2024a. An image is worth 1/2 tokens after layer 2: Plug-and-play inference acceleration for large vision-language models. In European Conference on Computer Vision, pages 19–35. Springer.
  9. Liang Chen, Haozhe Zhao, Tianyu Liu, Shuai Bai, Junyang Lin, Chang Zhou, and Baobao Chang. 2024b. An image is worth 1/2 tokens after layer 2: Plug-and-play inference acceleration for large vision-language models. In European Conference on Computer Vision, pages 19–35. Springer.
  10. Zeqi Chen, Zhaoyang Chu, Yi Gui, Feng Guo, Yao Wan, and Chuan Shi. 2025a. Bridging code graphs and large language models for better code understanding. arXiv preprint arXiv:2512.07666. — [arXiv:2512.07666](https://arxiv.org/abs/2512.07666) find
  11. Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, and 1 others. 2024c. Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling. arXiv preprint arXiv:2412.05271.
  12. Zhiwei Chen, Yupeng Hu, Zixu Li, Zhiheng Fu, Xuemeng Song, and Liqiang Nie. 2025b. Offset: Segmentation-based focus shift revision for composed image retrieval. In Proceedings of the ACM International Conference on Multimedia, page 6113–6122. find
  13. Kaichen Zhang* Fanyi Pu* Xinrun Du Yuhao Dong Haotian Liu Yuanhan Zhang Ge Zhang Chunyuan Li Bo Li*, Peiyuan Zhang* and Ziwei Liu. 2024. Lmms-eval: Accelerating the development of large multimoal models. find
  14. Wei-Lin Chiang, Zhuohan Li, Ziqing Lin, Ying Sheng, Zhanghao Wu, Hao Zhang, Lianmin Zheng, Siyuan Zhuang, Yonghao Zhuang, Joseph E Gonzalez, and 1 others. 2023. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality. See https://vicuna.lmsys. org (accessed 14 April 2023), 2(3):6. ↗
  15. Daniel Bolya, Cheng-Yang Fu, Xiaoliang Dai, Peizhao Zhang, Christoph Feichtenhofer, and Judy Hoffman. 2023. Token merging: Your vit but faster. In Proceedings of the International Conference on Learning Representations. find
  16. Ian Covert, Tony Sun, James Zou, and Tatsunori Hashimoto. 2024. Locality alignment improves vision-language models. arXiv preprint arXiv:2410.11087. — [arXiv:2410.11087](https://arxiv.org/abs/2410.11087) find
  17. T Brown, B Mann, N Ryder, M Subbiah, JD Kaplan, P Dhariwal, A Neelakantan, P Shyam, G Sastry, A Askell, and 1 others. 2020. Language models are few-shot learners advances in neural information processing systems 33.
  18. Mu Cai, Jianwei Yang, Jianfeng Gao, and Yong Jae Lee. 2024. Matryoshka multimodal models. In Workshop on Video-Language Models@ NeurIPS 2024. find
  19. Hugging Face. 2024. Mastering Transformers: The Journey from BERT to Large Language Models and Stable Diffusion. Open Library. find
  20. Chaoyou Fu, Peixian Chen, Yunhang Shen, Yulei Qin, Mengdan Zhang, Xu Lin, Zhenyu Qiu, Wei Lin, Jinrui Yang, Xiawu Zheng, Ke Li, Xing Sun, and Rongrong Ji. 2023. Mme: A comprehensive evaluation benchmark for multimodal large language models. ArXiv, abs/2306.13394.
  21. Julia Hirschberg and Christopher D Manning. 2015. Advances in natural language processing. Science, 349(6245):261–266. find
  22. E. Huang and 1 others. 2023. Evaluating large language models in complex scenarios. Journal of Computational Linguistics. find
  23. Zixu Li, Yupeng Hu, Zhiwei Chen, Qinlei Huang, Guozhi Qiu, Zhiheng Fu, and Meng Liu. 2026b. Retrack: Evidence-driven dual-stream directional anchor calibration network for composed video retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 23373–23381. find
  24. Kai Huang, Hao Zou, Ye Xi, BoChen Wang, Zhen Xie, and Liang Yu. 2024. Ivtp: Instruction-guided visual token pruning for large vision-language models. In European Conference on Computer Vision, pages 214–230. Springer. find
  25. Zixu Li, Yupeng Hu, Zhiwei Chen, Shiqi Zhang, Qinlei Huang, Zhiheng Fu, and Yinwei Wei. 2026c. Habit: Chrono-synergia robust progressive learning framework for composed image retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 6762–6770. find
  26. Feyza Duman Keles, Pruthuvi Mahesakya Wijewardena, and Chinmay Hegde. 2023. On the computational complexity of self-attention. In International conference on algorithmic learning theory, pages 597–619. PMLR. find
  27. Bin Lin, Yang Ye, Bin Zhu, Jiaxi Cui, Munan Ning, Peng Jin, and Li Yuan. 2023. Video-llava: Learning united visual representation by alignment before projection. arXiv preprint arXiv:2311.10122.
  28. Brett Koonce. 2021. Resnet 50. In Convolutional neural networks with swift for tensorflow: image recognition and dataset categorization, pages 63–72. Springer. find
  29. Zhihang Lin, Mingbao Lin, Luxi Lin, and Rongrong Ji. 2025a. Boosting multimodal large language models with visual tokens withdrawal for rapid inference. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 5334–5342. find
  30. Jaewoo Lee, Keyang Xuan, Chanakya Ekbote, Sandeep Polisetty, Yi R Fung, and Paul Pu Liang. 2025. Tamp: Token-adaptive layerwise pruning in multimodal large language models. arXiv preprint arXiv:2504.09897. find
  31. C. Li and 1 others. 2023a. Fine-tuning techniques for efficient model adaptation. AI Research Journal. find
  32. Chenglin Li, Qianglong Chen, Feng Han, Yikun Wang, Xingxi Yin, Yan Gong, Ruilin Li, Yin Zhang, and Jiaqi Wang. 2026a. Videothinker: Building agentic videollms with llm-guided tool reasoning. arXiv preprint arXiv:2601.15724. — [arXiv:2601.15724](https://arxiv.org/abs/2601.15724) find
  33. Duo Li, Zuhao Yang, and Shijian Lu. 2025a. Todre: Visual token pruning via diversity and task awareness for efficient large vision-language models. arXiv preprint arXiv:2505.18757. — [arXiv:2505.18757](https://arxiv.org/abs/2505.18757) find
  34. Mingxiao Li, Na Su, Fang Qu, Zhizhou Zhong, Ziyang Chen, Yuan Li, Zhaopeng Tu, and Xiaolong Li. 2025b. Vista: Enhancing vision-text alignment in mllms via cross-modal mutual information maximization. arXiv preprint arXiv:2505.10917. — [arXiv:2505.10917](https://arxiv.org/abs/2505.10917) find
  35. Wentong Li, Yuqian Yuan, Jian Liu, Dongqi Tang, Song Wang, Jie Qin, Jianke Zhu, and Lei Zhang. 2025c. Tokenpacker: Efficient visual projector for multimodal llm. International Journal of Computer Vision, pages 1–19. find
  36. Yanshu Li, Jianjiang Yang, Zhennan Shen, Ligong Han, Haoyan Xu, and Ruixiang Tang. 2025d. Catp: Contextually adaptive token pruning for efficient and enhanced multimodal in-context learning. arXiv preprint arXiv:2508.07871. — [arXiv:2508.07871](https://arxiv.org/abs/2508.07871) find
  37. Yifan Li, Yifan Du, Kun Zhou, Jinpeng Wang, Wayne Xin Zhao, and Ji-Rong Wen. 2023b. Evaluating object hallucination in large vision-language models. arXiv preprint arXiv:2305.10355.
  38. Zhihang Lin, Mingbao Lin, Luxi Lin, and Rongrong Ji. 2025b. Boosting multimodal large language models with visual tokens withdrawal for rapid inference. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 5334–5342. find
  39. Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. 2024a. Improved baselines with visual instruction tuning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26296–26306.
  40. Haotian Liu, Chunyuan Li, Yuheng Li, Bo Li, Yuanhan Zhang, Sheng Shen, and Yong Jae Lee. 2024b. Llavanext: Improved reasoning, ocr, and world knowledge. find
  41. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. Visual instruction tuning. Advances in neural information processing systems, 36:34892– 34916.
  42. Shizhan Liu, Hang Yu, Cong Liao, Jianguo Li, Weiyao Lin, Alex X Liu, and Schahram Dustdar. 2022. Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting. In \# PLACEHOLDER_PARENT_METADATA_VALUE\#. find
  43. Ting Liu, Liangtao Shi, Richang Hong, Yue Hu, Quanjun Yin, and Linfeng Zhang. 2024c. Multi-stage vision token dropping: Towards efficient multimodal large language model. arXiv preprint arXiv:2411.10803. — [arXiv:2411.10803](https://arxiv.org/abs/2411.10803) find
  44. Yuan Liu, Haodong Duan, Yuanhan Zhang, Bo Li, Songyang Zhang, Wangbo Zhao, Yike Yuan, Jiaqi Wang, Conghui He, Ziwei Liu, and 1 others. 2024d. Mmbench: Is your multi-modal model an all-around player? In European conference on computer vision, pages 216–233. Springer.
  45. Yuliang Liu, Zhang Li, Mingxin Huang, Biao Yang, Wenwen Yu, Chunyuan Li, Xu-Cheng Yin, Cheng-Lin Liu, Lianwen Jin, and Xiang Bai. 2024e. Ocrbench: on the hidden mystery of ocr in large multimodal models. Science China Information Sciences, 67(12):220102.
  46. Kenneth Marino, Mohammad Rastegari, Ali Farhadi, and Roozbeh Mottaghi. 2019. Ok-vqa: A visual question answering benchmark requiring external knowledge. In Proceedings of the IEEE/cvf conference on computer vision and pattern recognition, pages 3195–3204. find
  47. Ahmed Masry, Do Xuan Long, Jia Qing Tan, Shafiq Joty, and Enamul Hoque. 2022. Chartqa: A benchmark for question answering about charts with visual and logical reasoning. arXiv preprint arXiv:2203.10244.
  48. Mark EJ Newman. 2005. Power laws, pareto distributions and zipf’s law. Contemporary physics, 46(5):323–351. find
  49. Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, and 1 others. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32.
  50. Fabian Pedregosa, Gael Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, and 1 others. 2011. Scikit-learn: Machine learning in python journal of machine learning research. Journal of machine learning rese… find
  51. Ziheng Qin, Zhaopan Xu, Yukun Zhou, Zangwei Zheng, Zebang Cheng, Hao Tang, Lei Shang, Baigui Sun, Xiaojiang Peng, Radu Timofte, and 1 others. 2024. Dataset growth. In European Conference on Computer Vision, pages 58–75. Springer. find
  52. Yuzhang Shang, Mu Cai, Bingxin Xu, Yong Jae Lee, and Yan Yan. 2024a. Llava-prumerge: Adaptive token reduction for efficient large multimodal models. arXiv preprint arXiv:2403.15388. — [arXiv:2403.15388](https://arxiv.org/abs/2403.15388) find
  53. Yuzhang Shang, Mu Cai, Bingxin Xu, Yong Jae Lee, and Yan Yan. 2024b. Llava-prumerge: Adaptive token reduction for efficient large multimodal models. arXiv preprint arXiv:2403.15388. — [arXiv:2403.15388](https://arxiv.org/abs/2403.15388) find
  54. Piyush Sharma, Nan Ding, Sebastian Goodman, and Radu Soricut. 2018. Conceptual captions: A cleaned, hypernymed, image alt-text dataset for automatic image captioning. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2556–2565.
  55. Amanpreet Singh, Vivek Natarajan, Meet Shah, Yu Jiang, Xinlei Chen, Dhruv Batra, Devi Parikh, and Marcus Rohrbach. 2019. Towards vqa models that can read. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8317–8326. find
  56. Yizheng Sun, Yanze Xin, Hao Li, Chenghua Lin, and Riza Batista-Navarro. Velar: Vision-oriented language-attentive token reduction for multimodal large language models. find
  57. Yizheng Sun, Yanze Xin, Hao Li, Jingyuan Sun, Chenghua Lin, and Riza Batista-Navarro. 2025. Lvpruning: An effective yet simple language-guided vision token pruning approach for multi-modal large language models. arXiv preprint arXiv:2501.13652. find
  58. Hind Taud and Jean-Franccois Mas. 2017. Multilayer perceptron (mlp). In Geomatic approaches for modeling land change scenarios, pages 451–455. Springer. find
  59. Kimi Team, Angang Du, Bohong Yin, Bowei Xing, Bowen Qu, Bowen Wang, Cheng Chen, Chenlin Zhang, Chenzhuang Du, Chu Wei, and 1 others. 2025. Kimi-vl technical report. arXiv preprint arXiv:2504.07491.
  60. Luke Tierney and Joseph B Kadane. 1986. Accurate approximations for posterior moments and marginal densities. Journal of the american statistical association, 81(393):82–86. find
  61. Jintao Tong, Wenwei Jin, Pengda Qin, Anqi Li, Yixiong Zou, Yuhong Li, Yuhua Li, and Ruixuan Li. 2025. Flowcut: Rethinking redundancy via information flow for efficient vision-language models. arXiv preprint arXiv:2505.19536. — [arXiv:2505.19536](https://arxiv.org/abs/2505.19536) find
  62. H Touvron, T Lavril, G Izacard, X Martinet, MA Lachaux, T Lacroix, B Rozière, N Goyal, E Hambro, F Azhar, and 1 others. 2023. Open and efficient foundation language models. Preprint at arXiv. https://doi.org/10.48550/arXiv, 2302. — [doi:10.48550/arxiv](https://doi.org/10.48550/arxiv) ↗
  63. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. arXiv preprint.
  64. Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, and 1 others. 2024. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191.
  65. Tongxi Wang. 2026. Fbs: Modeling native parallel reading inside a transformer. arXiv preprint arXiv:2601.21708.
  66. Thomas Wolf. 2020. Transformers: State-of-the-art natural language processing. arXiv preprint arXiv:1910.03771.
  67. Junhao Xiao, Zhiyu Wu, Hao Lin, Yi Chen, Yahui Liu, Xiaoran Zhao, Zixu Wang, and Zejiang He. 2026. Not just what’s there: Enabling clip to comprehend negated visual descriptions without fine-tuning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 10978–10986. find
  68. Long Xing, Qidong Huang, Xiaoyi Dong, Jiajie Lu, Pan Zhang, Yuhang Zang, Yuhang Cao, Conghui He, Jiaqi Wang, Feng Wu, and 1 others. 2024. Pyramiddrop: Accelerating your large vision-language models via pyramid visual redundancy reduction. arXiv preprint arXiv:2410.17247. — [arXiv:2410.17247](https://arxiv.org/abs/2410.17247) find
  69. Senqiao Yang, Yukang Chen, Zhuotao Tian, Chengyao Wang, Jingyao Li, Bei Yu, and Jiaya Jia. 2025. Visionzip: Longer is better but not necessary in vision language models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 19792–19802. find
  70. Weihao Ye, Qiong Wu, Wenhao Lin, and Yiyi Zhou. 2025a. Fit and prune: Fast and training-free visual token pruning for multi-modal large language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 22128–22136. find
  71. Weihao Ye, Qiong Wu, Wenhao Lin, and Yiyi Zhou. 2025b. Fit and prune: Fast and training-free visual token pruning for multi-modal large language models. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 22128–22136. find
  72. Hongxu Yin, Arash Vahdat, Jose M Alvarez, Arun Mallya, Jan Kautz, and Pavlo Molchanov. 2022. A-vit: Adaptive tokens for efficient vision transformer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10809– 10818. find
  73. Li Yuan, Yunpeng Chen, Tao Wang, Weihao Yu, Yujun Shi, Zi-Hang Jiang, Francis EH Tay, Jiashi Feng, and Shuicheng Yan. 2021. Tokens-to-token vit: Training vision transformers from scratch on imagenet. In Proceedings of the IEEE/CVF international conference on computer vision, pages 558–567. find
  74. Qizhe Zhang, Aosong Cheng, Ming Lu, Zhiyong Zhuo, Minqi Wang, Jiajun Cao, Shaobo Guo, Qi She, and Shanghang Zhang. 2024a. [cls] attention is all you need for training-free visual token pruning: Make vlm inference faster. arXiv preprint arXiv:2412.01818. — [arXiv:2412.01818](https://arxiv.org/abs/2412.01818) find
  75. Qizhe Zhang, Mengzhen Liu, Lichen Li, Ming Lu, Yuan Zhang, Junwen Pan, Qi She, and Shanghang Zhang. 2025e. Beyond attention or similarity: Maximizing conditional diversity for token pruning in mllms. arXiv preprint arXiv:2506.10967. find
  76. Yuan Zhang, Chun-Kai Fan, Junpeng Ma, Wenzhao Zheng, Tao Huang, Kuan Cheng, Denis Gudovskiy, Tomoyuki Okuno, Yohei Nakata, Kurt Keutzer, and 1 others. 2024b. Sparsevlm: Visual token sparsification for efficient vision-language model inference. arXiv preprint arXiv:2410.04417. find
  77. Yuanhan Zhang, Bo Li, haotian Liu, Yong jae Lee, Liangke Gui, Di Fu, Jiashi Feng, Ziwei Liu, and Chunyuan Li. 2024c. Llava-next: A strong zero-shot video understanding model. find
  78. Li Zheng, Hao Fei, Ting Dai, Zuquan Peng, Fei Li, Huisheng Ma, Chong Teng, and Donghong Ji. 2025. Multi-granular multimodal clue fusion for meme understanding. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 26057– 26065. find
  79. Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Yuchen Duan, Hao Tian, Weijie Su, Jie Shao, and 1 others. 2025. Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models. arXiv preprint arXiv:2504.10479.
  80. Hao Zhang, Bo Huang, Zhenjia Li, Xi Xiao, Hui Yi Leong, Zumeng Zhang, Xinwei Long, Tianyang Wang, and Hao Xu. 2025a. Sensitivity-lora: Low-load sensitivity-based fine-tuning for large language models. arXiv preprint arXiv:2509.09119. find
  81. Hao Zhang, Zhenjia Li, Runfeng Bao, Yifan Gao, Xi Xiao, Heng Zhang, Shuyang Zhang, Bo Huang, Yuhang Wu, Tianyang Wang, and 1 others. 2025b. Hyperadalora: Accelerating lora rank allocation during training via hypernetworks without sacrificing performance. arXiv preprint arXiv:2510.02630.
  82. Hao Zhang, Mengsi Lyu, Zhuo Chen, Xingrun Xing, Yulong Ao, and Yonghua Lin. 2025c. Pdtrim: Targeted pruning for prefill-decode disaggregation in inference. arXiv preprint arXiv:2509.04467.
  83. Hao Zhang, Mengsi Lyu, Bo Huang, Yulong Ao, and Yonghua Lin. 2025d. Trimtokenator-lc: Towards adaptive visual token pruning for large multimodal models with long contexts. arXiv preprint arXiv:2512.22748. — [arXiv:2512.22748](https://arxiv.org/abs/2512.22748) find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhang-2026-trimtokenator-adaptive-visual-token-pruning/