{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,13]],"date-time":"2026-06-13T15:46:01Z","timestamp":1781365561537,"version":"3.54.1"},"reference-count":87,"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"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.01483","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"15982-15993","source":"Crossref","is-referenced-by-count":3,"title":["Training-Free and Adaptive Sparse Attention for Efficient Long Video Generation"],"prefix":"10.1109","author":[{"given":"Yifei","family":"Xia","sequence":"first","affiliation":[{"name":"Peking University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Suhan","family":"Ling","sequence":"additional","affiliation":[{"name":"Peking University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fangcheng","family":"Fu","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yujie","family":"Wang","sequence":"additional","affiliation":[{"name":"Peking University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Huixia","family":"Li","sequence":"additional","affiliation":[{"name":"ByteDance Seed"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xuefeng","family":"Xiao","sequence":"additional","affiliation":[{"name":"ByteDance Seed"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bin","family":"Cui","sequence":"additional","affiliation":[{"name":"Peking University"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Star attention: Efficient 11 m inference over long sequences","author":"Acharya","year":"2024","journal-title":"arXiv preprint"},{"key":"ref2","author":"Bahdanau","year":"2016","journal-title":"Neural machine translation by jointly learning to align and translate"},{"key":"ref3","article-title":"Longformer: The long-document transformer","author":"Beltagy","year":"2020","journal-title":"arXiv preprint"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW59228.2023.00484"},{"key":"ref5","article-title":"Token merging: Your vit but faster","author":"Bolya","year":"2022","journal-title":"arXiv preprint"},{"key":"ref6","author":"Brooks","year":"2024","journal-title":"Video generation models as world simulators"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i2.32192"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00974"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-022-3700-8"},{"key":"ref10","author":"Chen","year":"2024","journal-title":"Pixart- \u03b4: Fast and controllable image generation with latent consistency models"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i2.32205"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00468"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1712"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4231-5"},{"key":"ref15","article-title":"Cat pruning: Cluster-aware token pruning for text-to-image diffusion models","author":"Cheng","year":"2025","journal-title":"arXiv preprint"},{"key":"ref16","article-title":"Generating long sequences with sparse transformers","author":"Child","year":"2019","journal-title":"arXiv preprint"},{"key":"ref17","article-title":"OpenCV Contributors","year":"2025","journal-title":"Opencv: Open source computer vision library"},{"key":"ref18","article-title":"Flashattention-2: Faster attention with better parallelism and work partitioning","author":"Dao","year":"2023","journal-title":"arXiv preprint"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1189"},{"key":"ref20","first-page":"8780","article-title":"Diffusion models beat gans on image synthesis","volume-title":"Advances in neural information processing systems","volume":"34","author":"Dhariwal","year":"2021"},{"key":"ref21","article-title":"Pipefusion: Patch-level pipeline parallelism for diffusion transformers inference","author":"Fang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.00730"},{"key":"ref23","article-title":"Gptq: Accurate post-training quantization for generative pre-trained transformers","author":"Frantar","year":"2022","journal-title":"arXiv preprint"},{"key":"ref24","article-title":"Seerattention: Learning intrinsic sparse attention in your llms","author":"Gao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref25","volume-title":"Block Sparse Attention","author":"Guo","year":"2024"},{"key":"ref26","article-title":"Lm-infinite: Simple on-the-fly length generalization for large language models","author":"Han","year":"2023","journal-title":"arXiv preprint"},{"key":"ref27","article-title":"Latent video diffusion models for high-fidelity long video generation","author":"He","year":"2022","journal-title":"arXiv preprint"},{"key":"ref28","article-title":"Classifier-free diffusion guidance","author":"Ho","year":"2022","journal-title":"arXiv preprint"},{"key":"ref29","first-page":"6840","article-title":"Denoising diffusion probabilistic models","volume":"33","author":"Ho","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0628"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00482"},{"key":"ref32","article-title":"Cogvideo: Large-scale pretraining for text-to-video generation via transformers","author":"Hong","year":"2022","journal-title":"arXiv preprint"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00513"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02060"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/3662158.3662806"},{"key":"ref36","article-title":"Sale: Low-bit estimation for efficient sparse attention in longcontext 11 m prefilling","author":"Ji","year":"2025","journal-title":"arXiv preprint"},{"key":"ref37","article-title":"Minference 1.0: Accelerating pre-filling for long-context 11 ms via dynamic sparse attention","author":"Jiang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00639"},{"key":"ref39","article-title":"Adaptive caching for faster video generation with diffusion transformers","author":"Kahatapitiya","year":"2024","journal-title":"arXiv preprint"},{"key":"ref40","article-title":"Hunyuanvideo: A systematic framework for large video generative models","author":"Kong","year":"2024","journal-title":"arXiv preprint"},{"key":"ref41","article-title":"Black Forest Labs","volume-title":"Flux","year":"2024"},{"key":"ref42","article-title":"Svdqunat: Absorbing outliers by lowrank components for 4-bit diffusion models","author":"Li","year":"2024","journal-title":"arXiv preprint"},{"key":"ref43","article-title":"Open-sora plan: Open-source large video generation model","author":"Lin","year":"2024","journal-title":"arXiv preprint"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2786"},{"key":"ref45","first-page":"87","article-title":"Awq: Activation-aware weight quantization for on-device 11 m compression and acceleration","volume-title":"Proceedings of machine learning and systems","volume":"6","author":"Lin","year":"2024"},{"key":"ref46","article-title":"Faster diffusion via temporal attention decomposition","author":"Liu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref47","article-title":"Transformer acceleration with dynamic sparse attention","author":"Liu","year":"2021","journal-title":"arXiv preprint"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01492"},{"key":"ref49","article-title":"X-former elucidator: reviving efficient attention for long context language modeling","volume-title":"Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI)","author":"Miao","year":"2024"},{"key":"ref50","author":"Milakov","year":"2018","journal-title":"Online normalizer calculation for softmax"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00387"},{"key":"ref52","article-title":"Dreamfusion: Text-to-3d using 2d diffusion","author":"Poole","year":"2022","journal-title":"arXiv preprint"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72633-0_24"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1007\/s11390-024-4125-1"},{"key":"ref55","article-title":"Blockwise self-attention for long document understanding","author":"Qiu","year":"2019","journal-title":"arXiv preprint"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00710"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1145\/3680528.3687625"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2193"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00852"},{"key":"ref61","article-title":"Denoising diffusion implicit models","author":"Song","year":"2020","journal-title":"arXiv preprint"},{"key":"ref62","article-title":"Dsv: Exploiting dynamic sparsity to accelerate large-scale video dit training","author":"Tan","year":"2025","journal-title":"arXiv preprint"},{"key":"ref63","article-title":"Video-infinity: Distributed long video generation","author":"Tan","year":"2024","journal-title":"arXiv preprint"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1145\/3530811"},{"key":"ref65","article-title":"Genmo Team","volume-title":"Mochi 1","year":"2024"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1145\/3315508.3329973"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.spnlp-1.7"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1007\/s41019-023-00208-9"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2003.819861"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00701"},{"key":"ref72","article-title":"Sparse videogen: Accelerating video diffusion transformers with spatial-temporal sparsity","author":"Xi","year":"2025","journal-title":"arXiv preprint"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.52202\/079017-2034"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51701.2025.01483"},{"key":"ref75","first-page":"38087","article-title":"Smoothquant: Accurate and efficient post-training quantization for large language models","volume-title":"International conference on machine learning","author":"Xiao","year":"2023"},{"key":"ref76","first-page":"17453","volume-title":"Efficient streaming language models with attention sinks","author":"Xiao","year":"2023"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1007\/s41019-023-00226-7"},{"key":"ref78","article-title":"Cogvideox: Text-to-video diffusion models with an expert transformer","author":"Yang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i15.29599"},{"key":"ref80","first-page":"17283","article-title":"Big bird: Trans-formers for longer sequences","volume":"33","author":"Zaheer","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref81","doi-asserted-by":"publisher","DOI":"10.1007\/s11390-023-2386-8"},{"key":"ref82","article-title":"Sageattention2: Efficient attention with thorough outlier smoothing and per-thread int4 quantization","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref83","article-title":"Sageattention: Accurate 8-bit attention for plug-and-play inference acceleration","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref84","article-title":"Fast video generation with sliding tile attention","author":"Zhang","year":"2025","journal-title":"arXiv preprint"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00068"},{"key":"ref86","article-title":"Realtime video generation with pyramid attention broadcast","author":"Zhao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref87","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\/11445519.pdf?arnumber=11445519","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:32:15Z","timestamp":1777613535000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11445519\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":87,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.01483","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}