Cover Image

Attend to Not Attended: Structure-then-Detail Token Merging for Post-training DiT Acceleration

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Fang, Haipeng, Tang, Sheng, Cao, Juan, Zhang, Enshuo, Tang, Fan, Lee, Tong-Yee
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

HeadRouter: A Training-free Image Editing Framework for MM-DiTs by Adaptively Routing Attention Heads
by: Xu, Yu, et al.
Published: (2024)

PTQ4DiT: Post-training Quantization for Diffusion Transformers
by: Wu, Junyi, et al.
Published: (2024)

Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation
by: Chen, Ruoyu, et al.
Published: (2025)

Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers
by: Chen, Lei, et al.
Published: (2024)

Reasoning to Attend: Try to Understand How <SEG> Token Works
by: Qian, Rui, et al.
Published: (2024)

Locality-Attending Vision Transformer
by: Hajimiri, Sina, et al.
Published: (2026)

$Δ$-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers
by: Chen, Pengtao, et al.
Published: (2024)

Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformers
by: Ren, Sucheng, et al.
Published: (2025)

OUSAC: Optimized Guidance Scheduling with Adaptive Caching for DiT Acceleration
by: Sun, Ruitong, et al.
Published: (2025)

U-DiTs: Downsample Tokens in U-Shaped Diffusion Transformers
by: Tian, Yuchuan, et al.
Published: (2024)

DiVE: DiT-based Video Generation with Enhanced Control
by: Jiang, Junpeng, et al.
Published: (2024)

YOSE: You Only Select Essential Tokens for Efficient DiT-based Video Object Removal
by: Wu, Chenyang, et al.
Published: (2026)

HyperMotionX: The Dataset and Benchmark with DiT-Based Pose-Guided Human Image Animation of Complex Motions
by: Xu, Shuolin, et al.
Published: (2025)

Q-DiT4SR: Exploration of Detail-Preserving Diffusion Transformer Quantization for Real-World Image Super-Resolution
by: Zhang, Xun, et al.
Published: (2026)

FIS-DiT: Breaking the Few-Step Video Inference Barrier via Training-Free Frame Interleaved Sparsity
by: Tang, Jian, et al.
Published: (2026)

UniAnimate-DiT: Human Image Animation with Large-Scale Video Diffusion Transformer
by: Wang, Xiang, et al.
Published: (2025)

Inf-DiT: Upsampling Any-Resolution Image with Memory-Efficient Diffusion Transformer
by: Yang, Zhuoyi, et al.
Published: (2024)

Attend and Enrich: Enhanced Visual Prompt for Zero-Shot Learning
by: Liu, Man, et al.
Published: (2024)

Remix-DiT: Mixing Diffusion Transformers for Multi-Expert Denoising
by: Fang, Gongfan, et al.
Published: (2024)

LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Image and Video Generation
by: Yang, Lianwei, et al.
Published: (2025)

SparseDiT: Token Sparsification for Efficient Diffusion Transformer
by: Chang, Shuning, et al.
Published: (2024)

GATS: Gather-Attend-Scatter
by: Zolna, Konrad, et al.
Published: (2024)

Attend to what I say: Highlighting relevant content on slides
by: M, Megha Mariam K, et al.
Published: (2026)

DiTalker: A Unified DiT-based Framework for High-Quality and Speaking Styles Controllable Portrait Animation
by: Feng, He, et al.
Published: (2025)

DiT4Edit: Diffusion Transformer for Image Editing
by: Feng, Kunyu, et al.
Published: (2024)

LaVin-DiT: Large Vision Diffusion Transformer
by: Wang, Zhaoqing, et al.
Published: (2024)

DiT-Air: Revisiting the Efficiency of Diffusion Model Architecture Design in Text to Image Generation
by: Chen, Chen, et al.
Published: (2025)

Unveiling Redundancy in Diffusion Transformers (DiTs): A Systematic Study
by: Sun, Xibo, et al.
Published: (2024)

GeneralAD: Anomaly Detection Across Domains by Attending to Distorted Features
by: Sträter, Luc P. J., et al.
Published: (2024)

Perturb, Attend, Detect and Localize (PADL): Robust Proactive Image Defense
by: Bartolucci, Filippo, et al.
Published: (2024)

RealisDance-DiT: Simple yet Strong Baseline towards Controllable Character Animation in the Wild
by: Zhou, Jingkai, et al.
Published: (2025)

GenMask: Adapting DiT for Segmentation via Direct Mask Generation
by: Yang, Yuhuan, et al.
Published: (2026)

MaterialPicker: Multi-Modal DiT-Based Material Generation
by: Ma, Xiaohe, et al.
Published: (2024)

Insert Anything: Image Insertion via In-Context Editing in DiT
by: Song, Wensong, et al.
Published: (2025)

Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE
by: Shi, Yangming, et al.
Published: (2026)

ADP-DiT: Text-Guided Diffusion Transformer for Brain Image Generation in Alzheimer's Disease Progression
by: Lee, Juneyong, et al.
Published: (2026)

Lumina-Video: Efficient and Flexible Video Generation with Multi-scale Next-DiT
by: Liu, Dongyang, et al.
Published: (2025)

DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training
by: Feng, Haoran, et al.
Published: (2025)

Make-Your-Anchor: A Diffusion-based 2D Avatar Generation Framework
by: Huang, Ziyao, et al.
Published: (2024)

DSV: Exploiting Dynamic Sparsity to Accelerate Large-Scale Video DiT Training
by: Tan, Xin, et al.
Published: (2025)