{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T00:21:09Z","timestamp":1760487669185,"version":"build-2065373602"},"reference-count":14,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,9,22]],"date-time":"2025-09-22T00:00:00Z","timestamp":1758499200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,9,22]],"date-time":"2025-09-22T00:00:00Z","timestamp":1758499200000},"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":[[2025,9,22]]},"DOI":"10.1109\/icnp65844.2025.11192426","type":"proceedings-article","created":{"date-parts":[[2025,10,13]],"date-time":"2025-10-13T17:38:54Z","timestamp":1760377134000},"page":"1-6","source":"Crossref","is-referenced-by-count":0,"title":["Latency-Aware Transformer Partitioning for Heterogeneous Edge Inference"],"prefix":"10.1109","author":[{"given":"Xiaoyao","family":"Huang","sequence":"first","affiliation":[{"name":"Cloud Computing Research Institute,China Telecom,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref2","doi-asserted-by":"crossref","DOI":"10.1145\/3458817.3476209","article-title":"Efficient large-scale language model training on GPU clusters using megatron-lm","volume-title":"Proceedings of the 5th MLSys Conference","author":"Narayanan"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1145\/3093337.3037698"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2025.3562721"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/SWC62898.2024.00113"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/TITS.2025.3545126"},{"key":"ref7","article-title":"Dcha: Distributed-centralized heterogeneous architecture for multi-task smart sensing","author":"Shi","year":"2025","journal-title":"Design, Automation and Test in Europe (DATE)"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2024.3438995"},{"article-title":"Splitplace: Dynamic split computing for transformer models","volume-title":"NeurIPS Workshop on Efficient AI","author":"Kumar","key":"ref9"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM55648.2025.11044626"},{"key":"ref11","article-title":"Stressvit: Splitting and compressing transformers for edge-cloud collaboration","author":"Lin","year":"2024","journal-title":"Springer LNCS"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3672608.3707792"},{"article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","year":"2018","author":"Devlin","key":"ref13"},{"issue":"140","key":"ref14","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":"Journal of Machine Learning Research"}],"event":{"name":"2025 IEEE 33rd International Conference on Network Protocols (ICNP)","start":{"date-parts":[[2025,9,22]]},"location":"Seoul, Korea, Republic of","end":{"date-parts":[[2025,9,25]]}},"container-title":["2025 IEEE 33rd International Conference on Network Protocols (ICNP)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11192357\/11192322\/11192426.pdf?arnumber=11192426","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,14]],"date-time":"2025-10-14T05:12:34Z","timestamp":1760418754000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11192426\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,22]]},"references-count":14,"URL":"https:\/\/doi.org\/10.1109\/icnp65844.2025.11192426","relation":{},"subject":[],"published":{"date-parts":[[2025,9,22]]}}}