{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T10:14:56Z","timestamp":1777889696778,"version":"3.51.4"},"reference-count":60,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62325405,62104128,U19B2019,U21B2031,61832007,62204164"],"award-info":[{"award-number":["62325405,62104128,U19B2019,U21B2031,61832007,62204164"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100017582","name":"Beijing National Research Center for Information Science and Technology","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100017582","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01522","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"16399-16409","source":"Crossref","is-referenced-by-count":0,"title":["DiTFastAttnV2: Head-Wise Attention Compression for Multi-Modality Diffusion Transformers"],"prefix":"10.1109","author":[{"given":"Hanling","family":"Zhang","sequence":"first","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rundong","family":"Su","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhihang","family":"Yuan","sequence":"additional","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Pengtao","family":"Chen","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mingzhu","family":"Shen","sequence":"additional","affiliation":[{"name":"Imperial College London"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yibo","family":"Fan","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shengen","family":"Yan","sequence":"additional","affiliation":[{"name":"Infinigence AI"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Guohao","family":"Dai","sequence":"additional","affiliation":[{"name":"Infinigence AI"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yu","family":"Wang","sequence":"additional","affiliation":[{"name":"Tsinghua University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Gpt-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint"},{"key":"ref2","article-title":"Longformer: The long-document transformer","author":"Beltagy","year":"2020","journal-title":"arXiv preprint"},{"key":"ref3","article-title":"Stable video diffusion: Scaling latent video diffusion models to large datasets","author":"Blattmann","year":"2023","journal-title":"arXiv preprint"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW59228.2023.00484"},{"key":"ref5","volume-title":"Video generation models as world simulators","author":"Brooks","year":"2024"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW63382.2024.00087"},{"key":"ref7","article-title":"Pixart-: Fast training of diffusion transformer for photorealistic text-to-image synthesis","author":"Chen","year":"2023","journal-title":"arXiv preprint"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73411-3_5"},{"key":"ref9","article-title":"$\\delta$-dit: A training-free acceleration method tailored for diffusion transformers","author":"Chen","year":"2024","journal-title":"arXiv preprint"},{"key":"ref10","article-title":"Generating long sequences with sparse transformers","author":"Child","year":"2019","journal-title":"arXiv preprint"},{"key":"ref11","article-title":"Flashattention-2: Faster attention with better parallelism and work partitioning","author":"Dao","year":"2023","journal-title":"arXiv preprint"},{"key":"ref12","first-page":"8780","article-title":"Diffusion models beat gans on image synthesis","volume":"34","author":"Dhariwal","year":"2021","journal-title":"Advances in neural information processing systems"},{"key":"ref13","article-title":"An image is worth 16 16 words: Transformers for image recognition at scale","author":"Dosovitskiy","year":"2020","journal-title":"arXiv preprint"},{"key":"ref14","article-title":"Scaling rectified flow transformers for high-resolution image synthesis","volume-title":"Forty-first International Conference on Machine Learning","author":"Esser","year":"2024"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0731"},{"key":"ref16","article-title":"Ltx-video: Realtime video latent diffusion","author":"HaCohen","year":"2024","journal-title":"arXiv preprint"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00599"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0580"},{"key":"ref19","article-title":"Classifier-free diffusion guidance","author":"Ho","year":"2022","journal-title":"arXiv preprint"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1663"},{"key":"ref21","volume-title":"On architectural compression of text-toimage diffusion models","author":"Kim","year":"2023"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/WACV57701.2024.00141"},{"key":"ref23","article-title":"Hunyuanvideo: A systematic framework for large video generative models","author":"Kong","year":"2024","journal-title":"arXiv preprint"},{"key":"ref25","first-page":"arXiv-2312","article-title":"Faster diffusion: Rethinking the role of unet encoder in diffusion models","author":"Li","year":"2023","journal-title":"arXiv e-prints"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01608"},{"key":"ref27","article-title":"Clear: Conv-like linearization revs pre-trained diffusion transformers up","author":"Liu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01492"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref31","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"International conference on machine learning","author":"Radford","year":"2021"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"ref33","article-title":"Progressive distillation for fast sampling of diffusion models","author":"Salimans","year":"2022","journal-title":"arXiv preprint"},{"key":"ref34","article-title":"Improved techniques for training gans","volume":"29","author":"Salimans","year":"2016","journal-title":"Advances in neural information processing systems"},{"key":"ref35","article-title":"Fora: Fast-forward caching in diffusion transformer acceleration","author":"Selvaraju","year":"2024","journal-title":"arXiv preprint"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00196"},{"key":"ref37","article-title":"Todo: Token downsampling for efficient generation of high-resolution images","author":"Smith","year":"2024","journal-title":"arXiv preprint"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73464-9_20"},{"key":"ref39","article-title":"Generative modeling by estimating gradients of the data distribution","author":"Song","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref40","article-title":"Score-based generative modeling through stochastic differential equations","author":"Song","year":"2020","journal-title":"arXiv preprint"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73232-4_25"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.5555\/3157096.3157329"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00594"},{"key":"ref45","article-title":"Human preference score v 2: A solid benchmark for evaluating human preferences of text-to-image synthesis","author":"Wu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref46","article-title":"Sparse videogen: Accelerating video diffusion transformers with spatial-temporal sparsity","author":"Xi","year":"2025","journal-title":"arXiv preprint"},{"key":"ref47","article-title":"Efficient streaming language models with attention sinks","author":"Xiao","year":"2023","journal-title":"arXiv preprint"},{"key":"ref48","article-title":"Duoattention: Efficient long-context llm inference with retrieval and streaming heads","author":"Xiao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0700"},{"key":"ref50","article-title":"Cogvideox: Text-to-video diffusion models with an expert transformer","author":"Yang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref51","volume-title":"Ditfastattn: Attention compression for diffusion transformer models","author":"Yuan","year":"2024"},{"key":"ref52","article-title":"Laptop-diff: Layer pruning and normalized distillation for compressing diffusion models","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref53","article-title":"Sageattention2: Efficient attention with thorough outlier smoothing and per-thread int4 quantization","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref54","article-title":"Sageattention: Accurate 8-bit attention for plug-and-play inference acceleration","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref55","article-title":"Sageattention3: Microscaling fp4 attention for inference and an exploration of 8-bit training","author":"Zhang","year":"2025","journal-title":"arXiv preprint"},{"key":"ref56","article-title":"Fast video generation with sliding tile attention","author":"Zhang","year":"2025","journal-title":"arXiv preprint"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00068"},{"key":"ref58","article-title":"Crossattention makes inference cumbersome in text-to-image diffusion models","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref59","volume-title":"Initialization is critical to whether transformers fit composite functions by reasoning or memorizing","author":"Zhang"},{"key":"ref60","article-title":"Paroattention: Pattern-aware reordering for efficient sparse and quantized attention in visual generation models","author":"Zhao","year":"2025","journal-title":"arXiv preprint"},{"key":"ref61","volume-title":"Structured initialization for attention in vision transformers","author":"Zheng"},{"key":"ref62","article-title":"Open-sora: Democratizing efficient video production for all","author":"Zheng","year":"2024","journal-title":"arXiv preprint"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11445774.pdf?arnumber=11445774","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:16:55Z","timestamp":1777612615000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11445774\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":60,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01522","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}