{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,12]],"date-time":"2026-05-12T15:42:18Z","timestamp":1778600538637,"version":"3.51.4"},"reference-count":44,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"6","license":[{"start":{"date-parts":[[2025,6,1]],"date-time":"2025-06-01T00:00:00Z","timestamp":1748736000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,6,1]],"date-time":"2025-06-01T00:00:00Z","timestamp":1748736000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,6,1]],"date-time":"2025-06-01T00:00:00Z","timestamp":1748736000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Strategic Priority Research Program of the Chinese Academy of Sciences","award":["XDC05020100"],"award-info":[{"award-number":["XDC05020100"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Comput."],"published-print":{"date-parts":[[2025,6]]},"DOI":"10.1109\/tc.2025.3543697","type":"journal-article","created":{"date-parts":[[2025,2,19]],"date-time":"2025-02-19T18:41:32Z","timestamp":1739990492000},"page":"1857-1871","source":"Crossref","is-referenced-by-count":2,"title":["ETBench: Characterizing Hybrid Vision <u>T<\/u>ransformer Workloads Across <u>E<\/u>dge Devices"],"prefix":"10.1109","volume":"74","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-7847-7016","authenticated-orcid":false,"given":"Yingkun","family":"Zhou","sequence":"first","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-0745-2298","authenticated-orcid":false,"given":"Zhengshuyuan","family":"Tian","sequence":"additional","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-6511-4110","authenticated-orcid":false,"given":"Wenhao","family":"Yang","sequence":"additional","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1724-4904","authenticated-orcid":false,"given":"Tingting","family":"Zhang","sequence":"additional","affiliation":[{"name":"Loongson Technology Company Ltd., Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-7451-3200","authenticated-orcid":false,"given":"Jinpeng","family":"Ye","sequence":"additional","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-1247-9644","authenticated-orcid":false,"given":"Chenji","family":"Han","sequence":"additional","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5341-1343","authenticated-orcid":false,"given":"Tianyi","family":"Liu","sequence":"additional","affiliation":[{"name":"University of Texas at San Antonio, San Antonio, TX, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0430-3669","authenticated-orcid":false,"given":"Fuxin","family":"Zhang","sequence":"additional","affiliation":[{"name":"State Key Lab of Processors, Institute of Computing Technology, CAS, Beijing, China"}]}],"member":"263","reference":[{"key":"ref1","first-page":"770","article-title":"Deep residual learning for image recognition","volume-title":"Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR)","author":"He","year":"2015"},{"key":"ref2","first-page":"1314","article-title":"Searching for mobilenetv3","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. (ICCV)","author":"Howard","year":"2019"},{"key":"ref3","first-page":"10096","article-title":"Efficientnetv2: Smaller models and faster training","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Tan","year":"2021"},{"key":"ref4","first-page":"5998","article-title":"Attention is all you need","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Vaswani","year":"2017"},{"key":"ref5","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2020.acl-main.195","article-title":"MobileBERT: A compact task-agnostic BERT for resource-limited devices","volume-title":"Proc. Annu. Meeting Assoc. Comput. Linguistics","author":"Sun","year":"2020"},{"key":"ref6","article-title":"Cluecorpus2020: A large-scale Chinese corpus for pre-training language model","author":"Xu","year":"2020"},{"key":"ref7","article-title":"A first look at deep learning apps on smartphones","volume-title":"Proc. World Wide Web Conf.","author":"Xu","year":"2018"},{"key":"ref8","first-page":"331","article-title":"Machine learning at Facebook: Understanding inference at the edge","volume-title":"Proc. IEEE Int. Symp. High Perform. Comput. Archit. (HPCA)","author":"Wu","year":"2019"},{"key":"ref9","doi-asserted-by":"crossref","first-page":"658","DOI":"10.1145\/3487552.3487863","article-title":"Smart at what cost?: Characterising mobile deep neural networks in the wild","volume-title":"Proc. 21st ACM Internet Meas. Conf.","author":"Almeida","year":"2021"},{"key":"ref10","first-page":"1389","article-title":"Rethinking mobile block for efficient attention-based models","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. (ICCV)","author":"Zhang","year":"2023"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01549"},{"key":"ref12","first-page":"17379","article-title":"SwiftFormer: Efficient additive attention for transformer-based real-time mobile vision applications","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. (ICCV)","author":"Shaker","year":"2023"},{"key":"ref13","first-page":"12239","article-title":"LeViT: A vision transformer in ConvNet\u2019s clothing for faster inference","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. (ICCV)","author":"Graham","year":"2021"},{"key":"ref14","article-title":"EasyQuant: Post-training quantization via scale optimization","author":"Wu","year":"2020"},{"key":"ref15","article-title":"Tensorflow lite","year":"2024"},{"key":"ref16","first-page":"1","article-title":"MNN: A universal and efficient inference engine","volume-title":"Proc. MLSys","author":"Jiang"},{"key":"ref17","article-title":"Openvino","year":"2024"},{"key":"ref18","article-title":"Coreml","year":"2024"},{"key":"ref19","first-page":"1","article-title":"Nvidia Orin system-on-chip","volume-title":"IEEE Hot Chips 34 Symp. (HCS)","author":"Ditty","year":"2022"},{"key":"ref20","first-page":"789","article-title":"Ascend: A scalable and unified architecture for ubiquitous deep neural network computing: Industry track paper","volume-title":"Proc. IEEE Int. Symp. High-Perform. Comput. Archit. (HPCA)","author":"Liao","year":"2021"},{"key":"ref21","article-title":"Comparison and benchmarking of AI models and frameworks on mobile devices","author":"Luo","year":"2020"},{"key":"ref22","first-page":"3298","article-title":"A comprehensive benchmark of deep learning libraries on mobile devices","volume-title":"Proc. ACM Web Conf.","author":"Zhang","year":"2022"},{"key":"ref23","first-page":"3617","article-title":"Ai benchmark: All about deep learning on smartphones in 2019","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. Workshop (ICCVW)","author":"Ignatov","year":"2019"},{"key":"ref24","article-title":"MLPerf mobile inference benchmark: An industry-standard open-source machine learning benchmark for on-device AI","volume-title":"Proc. Conf. Mach. Learn. Syst.","author":"Reddi","year":"2020"},{"key":"ref25","article-title":"Separable self-attention for mobile vision transformers","author":"Mehta","year":"2022"},{"key":"ref26","article-title":"MobileViT: Light-weight, general-purpose, and mobile-friendly vision transformer","author":"Mehta","year":"2021"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-25082-8_1"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52688.2022.00520"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01386"},{"key":"ref30","article-title":"ARM A64 SIMD FP instruction set architecture manual","year":"2024"},{"key":"ref31","article-title":"Intel deep learning boost with vector neural network instructions (VNNI)","year":"2024"},{"key":"ref32","first-page":"578","article-title":"$\\{${TVM$\\}$}: An automated $\\{${end-to-end$\\}$} optimizing compiler for deep learning","volume-title":"Proc. 13th USENIX Symp. Operat. Syst. Des. Implement. (OSDI)","author":"Chen","year":"2018"},{"key":"ref33","first-page":"3389","article-title":"Learning to optimize tensor programs","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Chen","year":"2018"},{"key":"ref34","first-page":"863","article-title":"Ansor: Generating $\\{${high-performance$\\}$} tensor programs for deep learning","volume-title":"Proc. 14th USENIX Symp. Operat. Syst. Des. Implement. (OSDI)","author":"Zheng","year":"2020"},{"key":"ref35","article-title":"Mobile AI bench","year":"2024"},{"key":"ref36","article-title":"fvcore","year":"2024"},{"key":"ref37","article-title":"Perf","year":"2024"},{"key":"ref38","first-page":"4738","article-title":"Low-level optimizations for faster mobile deep learning inference frameworks","volume-title":"Proc. 28th ACM Int. Conf. Multimedia","author":"Febvay","year":"2020"},{"key":"ref39","first-page":"1","article-title":"Performance characterization of using quantization for DNN inference on edge devices","volume-title":"Proc. IEEE 7th Int. Conf. Fog Edge Comput. (ICFEC)","author":"Ahn","year":"2023"},{"key":"ref40","first-page":"211","article-title":"ImageNet large scale visual recognition challenge","volume-title":"Int. J. Comput. Vis.","volume":"115","author":"Russakovsky","year":"2014"},{"key":"ref41","first-page":"164245","article-title":"Performance evaluation of INT8 quantized inference on mobile GPUs","volume-title":"IEEE Access","volume":"9","author":"Kim","year":"2021"},{"key":"ref42","first-page":"10425","article-title":"Designing network design spaces","volume-title":"Proc. IEEE\/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR)","author":"Radosavovic","year":"2020"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/ISPASS.2014.6844459"},{"key":"ref44","first-page":"674","article-title":"An empirical study on deployment faults of deep learning based mobile applications","volume-title":"Proc. IEEE\/ACM 43rd Int. Conf. Softw. Eng. (ICSE)","author":"Chen","year":"2021"}],"container-title":["IEEE Transactions on Computers"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/12\/10994769\/10892343.pdf?arnumber=10892343","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,5,12]],"date-time":"2025-05-12T17:44:01Z","timestamp":1747071841000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10892343\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6]]},"references-count":44,"journal-issue":{"issue":"6"},"URL":"https:\/\/doi.org\/10.1109\/tc.2025.3543697","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":[[2025,6]]}}}