{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,6]],"date-time":"2026-06-06T16:50:35Z","timestamp":1780764635649,"version":"3.54.1"},"reference-count":69,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"10","license":[{"start":{"date-parts":[[2025,10,1]],"date-time":"2025-10-01T00:00:00Z","timestamp":1759276800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,10,1]],"date-time":"2025-10-01T00:00:00Z","timestamp":1759276800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,1]],"date-time":"2025-10-01T00:00:00Z","timestamp":1759276800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Guangdong S&T Programme","award":["2024B0101040007"],"award-info":[{"award-number":["2024B0101040007"]}]},{"DOI":"10.13039\/501100021171","name":"Basic and Applied Basic Research Foundation of Guangdong Province","doi-asserted-by":"publisher","award":["2023B1515120058"],"award-info":[{"award-number":["2023B1515120058"]}],"id":[{"id":"10.13039\/501100021171","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Guangzhou Basic and Applied Basic Research Program","award":["2024A04J6367"],"award-info":[{"award-number":["2024A04J6367"]}]},{"name":"Program for Guangdong Introducing Innovative and Entrepreneurial Teams","award":["2017ZT07X355"],"award-info":[{"award-number":["2017ZT07X355"]}]},{"name":"Guangzhou Municipal Joint Funding Project with Universities and Enterprises","award":["2024A03J0616"],"award-info":[{"award-number":["2024A03J0616"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. on Mobile Comput."],"published-print":{"date-parts":[[2025,10]]},"DOI":"10.1109\/tmc.2025.3574695","type":"journal-article","created":{"date-parts":[[2025,5,29]],"date-time":"2025-05-29T13:35:45Z","timestamp":1748525745000},"page":"10945-10962","source":"Crossref","is-referenced-by-count":2,"title":["Resource-Efficient Collaborative Edge Transformer Inference With Hybrid Model Parallelism"],"prefix":"10.1109","volume":"24","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-8867-0655","authenticated-orcid":false,"given":"Shengyuan","family":"Ye","sequence":"first","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-4408-2832","authenticated-orcid":false,"given":"Bei","family":"Ouyang","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4707-9492","authenticated-orcid":false,"given":"Jiangsu","family":"Du","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liekang","family":"Zeng","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, The Chinese University of Hong Kong, Ma Liu Shui, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-2802-2258","authenticated-orcid":false,"given":"Tianyi","family":"Qian","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenzhong","family":"Ou","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9745-4372","authenticated-orcid":false,"given":"Xiaowen","family":"Chu","sequence":"additional","affiliation":[{"name":"Hong Kong University of Science and Technology, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4894-5540","authenticated-orcid":false,"given":"Deke","family":"Guo","sequence":"additional","affiliation":[{"name":"National Key Laboratory of Information Systems Engineering, National University of Defense Technology, Changsha, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5315-3375","authenticated-orcid":false,"given":"Yutong","family":"Lu","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9943-6020","authenticated-orcid":false,"given":"Xu","family":"Chen","sequence":"additional","affiliation":[{"name":"School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018"},{"key":"ref2","article-title":"Improving language understanding by generative pre-training","author":"Radford","year":"2018"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/comst.2025.3527641"},{"key":"ref4","article-title":"Sasha: Creative goal-oriented reasoning in smart homes with large language models","author":"King","year":"2023"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3387941"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.3390\/vehicles7010011"},{"key":"ref7","first-page":"663","article-title":"{AlpaServe}: Statistical multiplexing with model parallelism for deep learning serving","volume-title":"Proc. USENIX Conf. Operating Syst. Des. Implementation","author":"Li"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3437801.3441578"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2918951"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-57959-7"},{"key":"ref11","article-title":"Personal LLM agents: Insights and survey about the capability, efficiency and security","author":"Li","year":"2024"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3498361.3538932"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2020.3042320"},{"key":"ref14","article-title":"Beyond efficiency: A systematic survey of resource-efficient large language models","author":"Bai","year":"2024"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3545008.3545015"},{"key":"ref16","article-title":"Raspberry Pi 4 model B specifications","year":"2024"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/3447993.3448625"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2019.2944584"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/3302424.3303950"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2858384"},{"issue":"8","key":"ref22","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI Blog"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"issue":"140","key":"ref24","first-page":"1","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"J. Mach. Learn. Res."},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929"},{"key":"ref26","article-title":"OPT: Open pre-trained transformer language models","author":"Zhang","year":"2022"},{"key":"ref27","first-page":"119","article-title":"Ekya: Continuous learning of video analytics models on edge compute servers","volume-title":"Proc. 19th Symp. Netw. Syst. Des. Implementation","author":"Bhardwaj"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/SC41404.2022.00051"},{"key":"ref29","first-page":"22941","article-title":"On-device training under 256kb memory","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Lin"},{"key":"ref30","first-page":"19","article-title":"Communication efficient distributed machine learning with the parameter server","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Li"},{"key":"ref31","first-page":"103","article-title":"GPipe: Efficient training of giant neural networks using pipeline parallelism","volume-title":"Proc. Int. Conf. Neural Inf. Process. Syst.","author":"Huang"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359646"},{"key":"ref33","article-title":"Megatron-LM: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2019"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.134"},{"key":"ref35","first-page":"1","article-title":"Efficient large-scale language model training on GPU clusters using megatron-LM","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal.","author":"Narayanan"},{"key":"ref36","article-title":"Horovod: Fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM52122.2024.10621342"},{"key":"ref38","article-title":"Pytorch","year":"2019"},{"key":"ref39","article-title":"LLaMA.cpp: Port of Facebook\u2019s LLaMA model in C\/C","author":"Gerganov","year":"2023"},{"key":"ref40","first-page":"1","article-title":"MNN: A universal and efficient inference engine","volume-title":"Proc. Mach. Learn. Syst.","author":"Jiang"},{"key":"ref41","article-title":"Tensorflow-lite","year":"2021"},{"key":"ref42","article-title":"DistilBERT, a distilled version of BERT: Smaller, faster, cheaper and lighter","author":"Sanh","year":"2019"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W18-5446"},{"key":"ref44","first-page":"3637","article-title":"Matching networks for one shot learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Vinyals"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/jiot.2024.3524255"},{"key":"ref46","article-title":"Jetson-nano","year":"2019"},{"key":"ref47","article-title":"Nvidia geforce GTX 1080 Ti graphics card","year":"2017"},{"key":"ref48","first-page":"135","article-title":"Serverlessllm: Low-latency serverless inference for large language models","volume-title":"Proc. 18th USENIX Symp. Operating Syst. Des. Implementation","author":"Fu"},{"key":"ref49","article-title":"Fast distributed inference serving for large language models","author":"Wu","year":"2023"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1145\/3560905.3568520"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1145\/3498361.3538948"},{"key":"ref52","first-page":"87","article-title":"AWQ: Activation-aware weight quantization for on-device LLM compression and acceleration","volume-title":"Proc. Mach. Learn. Syst.","author":"Lin","year":"2024"},{"key":"ref53","first-page":"21702","article-title":"LLM-pruner: On the structural pruning of large language models","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Ma"},{"key":"ref54","article-title":"EchoLM: Accelerating LLM serving with real-time knowledge distillation","author":"Yu","year":"2025"},{"key":"ref55","article-title":"Empowering 1000 tokens\/second on-device LLM prefilling with MLLM-NPU","author":"Xu","year":"2024"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1145\/3694715.3695964"},{"key":"ref57","article-title":"Powerinfer-2: Fast large language model inference on a smartphone","author":"Xue","year":"2024"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00024"},{"key":"ref59","article-title":"Reducing activation recomputation in large transformer models","volume-title":"Proc. Mach. Learn. Syst.","author":"Korthikanti"},{"key":"ref60","first-page":"521","article-title":"Orca: A distributed serving system for {Transformer-based} generative models","volume-title":"Proc. USENIX Conf. Operating Syst. Des. Implementation","author":"Yu"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.1145\/3636534.3649363"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1145\/3673038.3673043"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.23919\/DATE58400.2024.10546617"},{"key":"ref64","article-title":"ZeRO : Extremely efficient collective communication for giant model training","author":"Wang","year":"2023"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1145\/3447993.3483278"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1145\/3570361.3613277"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1145\/3570361.3592505"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507778"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA52012.2021.00049"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1145\/3627535.3638466"}],"container-title":["IEEE Transactions on Mobile Computing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/7755\/11154819\/11017462.pdf?arnumber=11017462","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,10]],"date-time":"2025-09-10T19:54:02Z","timestamp":1757534042000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11017462\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10]]},"references-count":69,"journal-issue":{"issue":"10"},"URL":"https:\/\/doi.org\/10.1109\/tmc.2025.3574695","relation":{},"ISSN":["1536-1233","1558-0660","2161-9875"],"issn-type":[{"value":"1536-1233","type":"print"},{"value":"1558-0660","type":"electronic"},{"value":"2161-9875","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10]]}}}