{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,18]],"date-time":"2025-11-18T09:33:29Z","timestamp":1763458409500,"version":"3.29.0"},"reference-count":22,"publisher":"IEEE","license":[{"start":{"date-parts":[[2024,9,16]],"date-time":"2024-09-16T00:00:00Z","timestamp":1726444800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,9,16]],"date-time":"2024-09-16T00:00:00Z","timestamp":1726444800000},"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":[[2024,9,16]]},"DOI":"10.1109\/socc62300.2024.10737793","type":"proceedings-article","created":{"date-parts":[[2024,11,5]],"date-time":"2024-11-05T18:29:41Z","timestamp":1730831381000},"page":"1-6","source":"Crossref","is-referenced-by-count":1,"title":["Exploring Approximation and Dataflow Co-Optimization for Scalable Transformer Inference Architecture on the Edge"],"prefix":"10.1109","author":[{"given":"Liu","family":"He","sequence":"first","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yujin","family":"Wang","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zongle","family":"Huang","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shupei","family":"Fan","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chen","family":"Tang","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shuyuan","family":"Zhang","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Luchang","family":"Lei","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Huazhong","family":"Yang","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yongpan","family":"Liu","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hongyang","family":"Jia","sequence":"additional","affiliation":[{"name":"Tsinghua University,Department of Electronic Engineering,Beijing,China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20077-9_1"},{"key":"ref2","first-page":"10421","article-title":"Bevfusion: A simple and robust lidar-camera fusion framework","volume":"35","author":"Liang","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref3","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018","journal-title":"arXiv preprint arXiv:1810.04805"},{"key":"ref4","article-title":"Llama: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint arXiv:2302.13971"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2010.11929"},{"key":"ref6","first-page":"13937","article-title":"Dynamicvit: Efficient vision transformers with dynamic token sparsification","volume":"34","author":"Rao","year":"2021","journal-title":"Advances in neural information processing systems"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA51647.2021.00018"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507738"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2023.3234893"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/APCCAS.2018.8605654"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS45731.2020.9180870"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/DAC18074.2021.9586134"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/ICCAD57390.2023.10323725"},{"key":"ref14","first-page":"16344","article-title":"Flashattention: Fast and memory-efficient exact attention with io-awareness","volume":"35","author":"Dao","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref15","article-title":"Flashdecoding++: Faster large language model inference on gpus","author":"Hong","year":"2023","journal-title":"arXiv preprint arXiv:2311.01282"},{"key":"ref16","article-title":"Roberta: A robustly optimized bert pretraining approach","author":"Liu","year":"2019","journal-title":"arXiv preprint arXiv:1907.11692"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W18-5446"},{"key":"ref18","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","volume-title":"International conference on machine learning. PMLR","author":"Touvron"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref21","first-page":"5506","article-title":"I-bert: Integer-only bert quantization","volume-title":"International conference on machine learning. PMLR","author":"Kim"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2022\/164"}],"event":{"name":"2024 IEEE 37th International System-on-Chip Conference (SOCC)","start":{"date-parts":[[2024,9,16]]},"location":"Dresden, Germany","end":{"date-parts":[[2024,9,19]]}},"container-title":["2024 IEEE 37th International System-on-Chip Conference (SOCC)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10737725\/10737709\/10737793.pdf?arnumber=10737793","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,11,27]],"date-time":"2024-11-27T08:11:45Z","timestamp":1732695105000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10737793\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,9,16]]},"references-count":22,"URL":"https:\/\/doi.org\/10.1109\/socc62300.2024.10737793","relation":{},"subject":[],"published":{"date-parts":[[2024,9,16]]}}}