arXiv Computer Vision October 12, 2026 By Magdalena Tr\k{e}dowicz, Aleksandra Nowak, Tomasz Steifer, {\L}ukasz Struski, Szymon P{\l}otka, Jacek Tabor

LapMerge: Differentiable Exact-Budget Token Merging in Linear Time

Read the original on arXiv Computer Vision →

LapMerge is a new differentiable token merging operator that partitions an ordered sequence into exactly N equal‑mass segments using a learned Laplace‑kernel density, then aggregates each segment with prefix/suffix scans, achieving an exact output budget in linear time. The method preserves information while allowing end‑to‑end differentiability and can change the budget without retraining. Experiments on vision and vision‑language tasks show LapMerge matches or outperforms existing reducers and can be up to two orders of magnitude faster for long sequences, especially when token order reflects spatial or temporal locality.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

arXiv AI

Jun 16

Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality

arXiv:2505. 18227v4 Announce Type: replace-cross Abstract: In Transformer architectures, tokens\textemdash discrete units derived from raw data\textemdash are formed by segmenting inputs into fixed-length chunks.

By Zhenglun Kong, Yize Li, Fanhu Zeng, Lei Xin, Shvat Messica, Xue Lin, Pu Zhao, Manolis Kellis, Hao Tang, Marinka Zitnik

llms rag agents reinforcement-learning efficiency multimodal safety

arXiv Machine Learning

Sep 30

Beyond Selection: Token Parameterization for Extreme Visual Token Compression

arXiv:2609.35232v2 Announce Type: replace-cross Abstract: Visual-token compression is effective for improving the efficiency of vision-language models, but under extreme compression budgets, token pr...

By Rui Zhong, Yu Li, Zheyu Yan, Cheng Zhuo

llms rag efficiency multimodal safety

arXiv Machine Learning

5d ago

V-CoLA: Vision Token Compression with Linear Attention

The paper introduces V-CoLA, a training‑free token compression framework tailored for vision-language models that use linear attention. V-CoLA employs a uniqueness‑aware importance criterion to select critical vision tokens and an adaptive token merging strategy to compress them, all optimized to fit the chunk‑wise parallelism of linear attention. Experiments show that V‑CoLA retains 99.5% of the original performance with only half the vision tokens and achieves significant speedups during prefill.

By Hao Jiang, Yiru Mao, Tianpeng Bu, Hao Zhou, Hongtao Duan, Wang Jing, Bowen Xu, Xin Chen, Lulu Hu, Bin Yang, Yongliang Tao, Minying Zhang

llms multimodal benchmarks

arXiv Machine Learning

Aug 10

An AI4AI Framework for Visual Token Pruning

arXiv:2608. 07193v1 Announce Type: new Abstract: Visual-token pruning can substantially reduce the inference cost of multimodal large language models (MLLMs), yet existing methods largely rely on fixed, handcrafted heuristics and costly expert trial and error.

By Zhen Liu, Wenli Huang, Wei Song, Yuhan Liu, Zhiqin Yang, Jingwen Fu

llms efficiency multimodal benchmarks

Hugging Face Trending Papers

Aug 3

CRAFT: Compression via Recursive Adaptive Fusion of Video Tokens for Vision-Language Models

In video understanding, vision-language models (VLMs) must ingest massive numbers of visual tokens, causing the computational and memory cost of the prefill stage to rise sharply. Such visual sequences are highly redundant along the spatio-temporal dimension, yet a high compression ratio is often accompanied by the loss of critical details.

llms multimodal benchmarks safety

arXiv AI

Jun 19

Variable-Length Tokenization via Learnable Global Merging for Diffusion Transformers

arXiv:2606. 20076v1 Announce Type: cross Abstract: Latent Diffusion Models (LDMs) have become dominant in visual synthesis, but their quality-compute trade-off is largely constrained by the tokenizer's fixed compression ratio.

By Dong Hoon Lee, Seunghoon Hong

llms diffusion nlp safety

One email a morning, machine-written

Related stories

Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality

Beyond Selection: Token Parameterization for Extreme Visual Token Compression

V-CoLA: Vision Token Compression with Linear Attention

An AI4AI Framework for Visual Token Pruning

CRAFT: Compression via Recursive Adaptive Fusion of Video Tokens for Vision-Language Models

Variable-Length Tokenization via Learnable Global Merging for Diffusion Transformers