arXiv Computer Vision October 16, 2026 By SeungJu Cha, Ye-Chan Kim, HyunGee Kim, Sungho Koh, Dong-Jin Kim

Which Tokens to Merge? Diffusion Dynamics for Efficient Image Generation

Read the original on arXiv Computer Vision →

The paper introduces Dynamics-aware Token Merging (DaTMe), a training‑free method that accelerates diffusion models by selectively merging tokens based on diffusion dynamics. DaTMe builds an importance map from the Tweedie denoised estimate, combining temporal changes and structural high‑frequency content to decide which tokens to preserve and which to merge. Experiments on PixArt-a and FLUX show that DaTMe maintains visual fidelity across the entire image while reducing token computation.

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.

diffusion

arXiv Computer Vision

Aug 26

AffineTok: Semantic Affine Consistency for Diffusion-Friendly Visual Tokenizer

arXiv:2608.23864v1 Announce Type: new Abstract: Visual tokenizers increasingly inject semantic supervision into latent spaces to make downstream diffusion easier. Yet how these semantics should be or...

By Junqiu Yu, Pandeng Li, Yikai Wang, Jiaxing Zhao, Yujie Wei, Kaixun Jiang, Quanhao Li, Hongtao Yu, Zhihang Liu, Zhaohe Liao, Junjie Zhou, Yun Zheng, Yu Liu, Yanwei Fu

diffusion nlp benchmarks safety

arXiv Computer Vision

Aug 26

Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation

arXiv:2608.24293v1 Announce Type: new Abstract: Latent diffusion models have emerged as a dominant framework for high-fidelity image and video synthesis, operating in compact latent spaces with varia...

By Yeonkyeong Lee, Hyunsung Go, Jongmin Kim, Sewoong Lim, Donghoon Lee

llms diffusion nlp benchmarks

arXiv AI

Aug 5

Rectify Then Diffuse: Disentangling Concepts Before Denoising Trajectory Unfolds

arXiv:2608. 03135v1 Announce Type: cross Abstract: Text-to-image diffusion models can generate individual concepts well, but they often omit or merge concepts incorrectly with multiple concepts.

By Ning Zhu, An Chen, Mengfei Zhao, Juntao Xu, Jingze Liang, Boyuan Gu, Liang-Jian Deng

diffusion benchmarks

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 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

arXiv Machine Learning

Aug 4

DAVET: Denoising-Aware Visual Evidence Trajectory Allocation for Diffusion Vision-Language Models

arXiv:2608. 01821v1 Announce Type: cross Abstract: Diffusion vision-language models (dVLMs) iteratively denoise masked responses while conditioning each denoising step on visual evidence, making visual conditioning a substantial recurring inference cost.

By Yongkang Zhou, Xiang Xia, Cheng Yan, Fan Xu, Wuyang Zhang

llms diffusion efficiency multimodal benchmarks

One email a morning, machine-written

Related stories

AffineTok: Semantic Affine Consistency for Diffusion-Friendly Visual Tokenizer

Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation

Rectify Then Diffuse: Disentangling Concepts Before Denoising Trajectory Unfolds

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

Variable-Length Tokenization via Learnable Global Merging for Diffusion Transformers

DAVET: Denoising-Aware Visual Evidence Trajectory Allocation for Diffusion Vision-Language Models