{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,30]],"date-time":"2026-08-30T21:50:10Z","timestamp":1788126610483,"version":"build-2803163510"},"reference-count":37,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"7","license":[{"start":{"date-parts":[[2024,7,1]],"date-time":"2024-07-01T00:00:00Z","timestamp":1719792000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2024,7,1]],"date-time":"2024-07-01T00:00:00Z","timestamp":1719792000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,7,1]],"date-time":"2024-07-01T00:00:00Z","timestamp":1719792000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"publisher","award":["2023YFE0205700"],"award-info":[{"award-number":["2023YFE0205700"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62341410"],"award-info":[{"award-number":["62341410"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100003009","name":"Science and Technology Development Fund","doi-asserted-by":"publisher","award":["Macao S.A.R. (FDCT) Project 0078\/2023\/AMJ"],"award-info":[{"award-number":["Macao S.A.R. (FDCT) Project 0078\/2023\/AMJ"]}],"id":[{"id":"10.13039\/501100003009","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput."],"published-print":{"date-parts":[[2024,7]]},"DOI":"10.1109\/tc.2024.3389507","type":"journal-article","created":{"date-parts":[[2024,4,16]],"date-time":"2024-04-16T13:19:05Z","timestamp":1713273545000},"page":"1852-1865","source":"Crossref","is-referenced-by-count":12,"title":["Raptor-T: A Fused and Memory-Efficient Sparse Transformer for Long and Variable-Length Sequences"],"prefix":"10.1109","volume":"73","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-8329-0145","authenticated-orcid":false,"given":"Hulin","family":"Wang","sequence":"first","affiliation":[{"name":"School of Computer Science, Wuhan University, Hubei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3913-3623","authenticated-orcid":false,"given":"Donglin","family":"Yang","sequence":"additional","affiliation":[{"name":"Nvidia Corp., Santa Clara, CA, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-8101-785X","authenticated-orcid":false,"given":"Yaqi","family":"Xia","sequence":"additional","affiliation":[{"name":"School of Computer Science, Wuhan University, Hubei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6599-9976","authenticated-orcid":false,"given":"Zheng","family":"Zhang","sequence":"additional","affiliation":[{"name":"School of Computer Science, Wuhan University, Hubei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-8533-4726","authenticated-orcid":false,"given":"Qigang","family":"Wang","sequence":"additional","affiliation":[{"name":"AI Lab, Lenovo Research, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2290-1785","authenticated-orcid":false,"given":"Jianping","family":"Fan","sequence":"additional","affiliation":[{"name":"AI Lab, Lenovo Research, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-9500-3390","authenticated-orcid":false,"given":"Xiaobo","family":"Zhou","sequence":"additional","affiliation":[{"name":"IOTSC & Department of Computer and Information Sciences, University of Macau, Macau S.A.R., China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2869-7623","authenticated-orcid":false,"given":"Dazhao","family":"Cheng","sequence":"additional","affiliation":[{"name":"School of Computer Science, Wuhan University, Hubei, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"key":"ref3","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Proc. Adv. Neural Inf. Process. Syst. (NeurIPS)","author":"Brown","year":"2020"},{"key":"ref4","article-title":"Recurrent neural network regularization","author":"Zaremba","year":"2014"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"ref6","article-title":"DocBERT: BERT for document classification","author":"Adhikari","year":"2019"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.acl-short.79"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-1171"},{"key":"ref9","first-page":"1","article-title":"ELECTRA: Pre-training text encoders as discriminators rather than generators","volume-title":"Proc. 8th Int. Conf. Learn. Representations (ICLR)","author":"Clark","year":"2020"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-1412"},{"key":"ref11","first-page":"11328","article-title":"PEGASUS: Pre-training with extracted gap-sentences for abstractive summarization","volume-title":"Proc. 37th Int. Conf. Mach. Learn. (ICML)","author":"Zhang","year":"2020"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.252631999"},{"key":"ref13","article-title":"BP-Transformer: Modelling long-range context via binary partitioning","author":"Ye","year":"2019"},{"key":"ref14","article-title":"Longformer: The long-document transformer","author":"Beltagy","year":"2020"},{"key":"ref15","first-page":"1","article-title":"Reformer: The efficient transformer","volume-title":"Proc. 8th Int. Conf. Learn. Representations (ICLR)","author":"Kitaev","year":"2020"},{"key":"ref16","first-page":"17283","article-title":"Big bird: Transformers for longer sequences","volume-title":"Proc. Adv. Neural Inf. Process. Syst. (NeurIPS)","author":"Zaheer","year":"2020"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1145\/3437801.3441578"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS54959.2023.00042"},{"key":"ref19","first-page":"16344","article-title":"FlashAttention: Fast and memory-efficient exact attention with IO-awareness","volume-title":"Proc. Adv. Neural Inf. Process. Syst. (NeurIPS)","author":"Dao","year":"2022"},{"key":"ref20","first-page":"8024","article-title":"PyTorch: An imperative style, high-performance deep learning library","volume-title":"Proc. Adv. Neural Inf. Process. Syst. (NeurIPS)","author":"Paszke","year":"2019"},{"key":"ref21","article-title":"FasterTransformer","year":"2023"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"key":"ref23","article-title":"FlashAttention-2: Faster attention with better parallelism and work partitioning","author":"Dao","year":"2023"},{"key":"ref24","article-title":"Nvidia a100 tensor core GPU architecture","year":"2020"},{"key":"ref25","article-title":"WMMA","year":"2023"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/SC41404.2022.00043"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00024"},{"key":"ref28","article-title":"Megatron-lm: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2019"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.emnlp-demos.6"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1264"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P18-2124"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507738"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/SC41404.2022.00051"},{"key":"ref34","article-title":"TensorRT","year":"2023"},{"key":"ref35","first-page":"25","article-title":"ET: Re-thinking self-attention for transformer models on GPUs","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal. (SC)","author":"Chen","year":"2021"},{"key":"ref36","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","volume-title":"Proc. 12th USENIX Symp. Operating Syst. Des. Implementation (OSDI)","author":"Abadi","year":"2016"},{"key":"ref37","first-page":"578","article-title":"TVM: An automated end-to-end optimizing compiler for deep learning","volume-title":"Proc. 13th USENIX Symp. Operating Syst. Des. Implementation (OSDI)","author":"Chen","year":"2018"}],"container-title":["IEEE Transactions on Computers"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/12\/10552674\/10500743.pdf?arnumber=10500743","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,6,25]],"date-time":"2024-06-25T15:38:55Z","timestamp":1719329935000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10500743\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,7]]},"references-count":37,"journal-issue":{"issue":"7"},"URL":"https:\/\/doi.org\/10.1109\/tc.2024.3389507","relation":{},"ISSN":["0018-9340","1557-9956","2326-3814"],"issn-type":[{"value":"0018-9340","type":"print"},{"value":"1557-9956","type":"electronic"},{"value":"2326-3814","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,7]]}}}