{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,27]],"date-time":"2026-03-27T06:57:58Z","timestamp":1774594678394,"version":"3.50.1"},"reference-count":53,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,9,1]],"date-time":"2025-09-01T00:00:00Z","timestamp":1756684800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,9,1]],"date-time":"2025-09-01T00:00:00Z","timestamp":1756684800000},"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":[[2025,9,1]]},"DOI":"10.1109\/fpl68686.2025.00044","type":"proceedings-article","created":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T19:48:24Z","timestamp":1774554504000},"page":"263-272","source":"Crossref","is-referenced-by-count":0,"title":["Refining Datapath for Microscaling ViTs"],"prefix":"10.1109","author":[{"given":"Can","family":"Xiao","sequence":"first","affiliation":[{"name":"Imperial College London"}]},{"given":"Jianyi","family":"Cheng","sequence":"additional","affiliation":[{"name":"University of Edinburgh"}]},{"given":"Yiren","family":"Zhao","sequence":"additional","affiliation":[{"name":"Imperial College London"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3626202.3637562"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA56546.2023.10071047"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/FPL57034.2022.00027"},{"key":"ref4","author":"Alexey","year":"2020","journal-title":"An image is worth 16 \u00d7 16 words: Transformers for image recognition at scale"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/FPL64840.2024.00044"},{"key":"ref6","first-page":"10271","article-title":"Pushing the limits of narrow precision inferencing at cloud scale with microsoft floating point","volume":"33","author":"Darvish Rouhani","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.617"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/FPL64840.2024.00049"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2023.3312775"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20053-3_30"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"issue":"94720\u20131776","key":"ref12","first-page":"11","article-title":"Ieee standard 754 for binary floating-point arithmetic","volume":"754","author":"Kahan","year":"1996","journal-title":"Lecture Notes on the Status of IEEE"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3656177"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/FCCM.2016.22"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021744"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3617232.3624850"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/DAC56929.2023.10247797"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2024.3488966"},{"key":"ref19","year":"2025","journal-title":"PyTorch LayerNorm"},{"key":"ref20","year":"2025","journal-title":"Py Torch GELU"},{"key":"ref21","year":"2025","journal-title":"PyTorch ReLu6"},{"key":"ref22","year":"2025","journal-title":"PyTorch Softmax"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01565"},{"key":"ref24","year":"2025","journal-title":"llama.cpp"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0396"},{"key":"ref26","year":"2023","journal-title":"Pytorch Image Models"},{"key":"ref27","year":"2025","journal-title":"AMD Alveo V80 Compute Accelerator"},{"key":"ref28","author":"Lin","year":"2017","journal-title":"Towards accurate binary convolutional neural network"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01237-3_23"},{"key":"ref30","author":"Wu","year":"2018","journal-title":"Training and inference with integers in deep neural networks"},{"key":"ref31","author":"Krishnamoorthi","year":"2018","journal-title":"Quantizing deep convolutional networks for efficient inference: A whitepaper"},{"key":"ref32","first-page":"873","article-title":"Vs-quant: Per-vector scaled quantization for accurate low-precision neural network inference","volume-title":"Proceedings of Machine Learning and Systems","volume":"3","author":"Dai","year":"2021"},{"key":"ref33","author":"Harma","year":"2022","journal-title":"Accuracy boosters: Epoch-driven mixed-mantissa block floating-point for dnn training"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3579371.3589351"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO56248.2022.00048"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA45697.2020.00086"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1145\/3470496.3527438"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA52012.2021.00061"},{"key":"ref39","author":"Dettmers","year":"2022","journal-title":"Llm. int8 (): 8-bit matrix multiplication for transformers at scale"},{"key":"ref40","author":"Frantar","year":"2022","journal-title":"Gptq: Accurate post-training quantization for generative pre-trained transformers"},{"key":"ref41","author":"Xiao","year":"2022","journal-title":"Smoothquant: Accurate and efficient post-training quantization for large language models"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1970"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD58817.2023.00084"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA57654.2024.00082"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/FCCM57271.2023.00019"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO56248.2022.00050"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA47549.2020.00035"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO56248.2022.00051"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575747"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1145\/3466752.3480125"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00071"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1145\/3466752.3480095"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1145\/3579371.3589057"}],"event":{"name":"2025 35th International Conference on Field-Programmable Logic and Applications (FPL)","location":"Leiden, Netherlands","start":{"date-parts":[[2025,9,1]]},"end":{"date-parts":[[2025,9,5]]}},"container-title":["2025 35th International Conference on Field-Programmable Logic and Applications (FPL)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11449056\/11449057\/11449093.pdf?arnumber=11449093","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,27]],"date-time":"2026-03-27T05:27:12Z","timestamp":1774589232000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11449093\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,1]]},"references-count":53,"URL":"https:\/\/doi.org\/10.1109\/fpl68686.2025.00044","relation":{},"subject":[],"published":{"date-parts":[[2025,9,1]]}}}