{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,5]],"date-time":"2026-07-05T11:09:28Z","timestamp":1783249768429,"version":"3.54.6"},"reference-count":94,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key R&D Program of China","doi-asserted-by":"publisher","award":["2022ZD0161100"],"award-info":[{"award-number":["2022ZD0161100"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.00400","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"4198-4210","source":"Crossref","is-referenced-by-count":1,"title":["GenieBlue: Integrating Both Linguistic and Multimodal Capabilities for Large Language Models on Mobile Devices"],"prefix":"10.1109","author":[{"given":"Xudong","family":"Lu","sequence":"first","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yinghao","family":"Chen","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Renshou","family":"Wu","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haohao","family":"Gao","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xi","family":"Chen","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xue","family":"Yang","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiangyu","family":"Zhao","sequence":"additional","affiliation":[{"name":"Shanghai Jiao Tong University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aojun","family":"Zhou","sequence":"additional","affiliation":[{"name":"CUHK MMLab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fangyuan","family":"Li","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yafei","family":"Wen","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiaoxin","family":"Chen","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shuai","family":"Ren","sequence":"additional","affiliation":[{"name":"vivo AI Lab"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongsheng","family":"Li","sequence":"additional","affiliation":[{"name":"CUHK MMLab"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Phi-3 technical report: A highly capable language model locally on your phone","author":"Abdin","year":"2024","journal-title":"arXiv preprint"},{"key":"ref2","first-page":"23572367","article-title":"MathQA: Towards interpretable math word problem solving with operation-based formalisms","volume-title":"Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","volume":"1","author":"Amini"},{"key":"ref3","article-title":"Gemini: A family of highly capable multimodal models","volume":"1","author":"Anil","year":"2023","journal-title":"arXiv preprint"},{"key":"ref4","article-title":"Claude 3","year":"2023","journal-title":"Large Language Model."},{"key":"ref5","article-title":"Program synthesis with large language models","author":"Austin","year":"2021","journal-title":"arXiv preprint"},{"key":"ref6","article-title":"Openflamingo: An opensource framework for training large autoregressive visionlanguage models","author":"Awadalla","year":"2023","journal-title":"arXiv preprint"},{"key":"ref7","article-title":"Qwen2.5-vl technical report","author":"Bai","year":"2025","journal-title":"arXiv preprint"},{"key":"ref8","article-title":"Condor: Enhance 11 m alignment with knowledge-driven data synthesis and refinement","author":"Cao","year":"2025","journal-title":"arXiv preprint"},{"key":"ref9","volume-title":"openai summarize tldr dataset","year":"2023"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0850"},{"key":"ref11","article-title":"Evaluating large language models trained on code","author":"Chen","year":"2021","journal-title":"arXiv preprint"},{"key":"ref12","article-title":"Expanding performance boundaries of open-source multimodal models with model, data, and testtime scaling","author":"Chen","year":"2024","journal-title":"arXiv preprint"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4231-5"},{"key":"ref14","article-title":"Mobilevlm: A fast, reproducible and strong vision language assistant for mobile devices","author":"Chu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref15","article-title":"Mobilevlm v2: Faster and stronger baseline for vision language model","author":"Chu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref16","volume-title":"Think you have solved question answering? try arc, the ai2 reasoning challenge","author":"Clark","year":"2018"},{"key":"ref17","article-title":"Training verifiers to solve math word problems","author":"Cobbe","year":"2021","journal-title":"arXiv preprint"},{"key":"ref18","article-title":"Training verifiers to solve math word problems","author":"Cobbe","year":"2021","journal-title":"arXiv preprint"},{"key":"ref19","volume-title":"Free dolly: Introducing the world\u2019s first truly open instruction-tuned","volume":"11","author":"Conover","year":"2023"},{"key":"ref20","volume-title":"OpenCompass: A universal evaluation platform for foundation models","year":"2023"},{"key":"ref21","volume-title":"Grok-1.5 vision preview: Connecting the digital and physical worlds with our first multimodal model","year":"2024"},{"key":"ref22","article-title":"Ultrafeedback: Boosting language models with high-quality feedback","author":"Cui","year":"2023","journal-title":"arXiv preprint"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.183"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.1595"},{"key":"ref25","article-title":"Drop: A reading comprehension benchmark requiring discrete reasoning over paragraphs","author":"Dua","year":"2019","journal-title":"arXiv preprint"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1007\/s44267-024-00067-6"},{"key":"ref27","volume-title":"Glaive code assistant v3 dataset","year":"2024"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.01363"},{"key":"ref29","article-title":"Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning","author":"Guo","year":"2025","journal-title":"arXiv preprint"},{"key":"ref30","article-title":"Measuring massive multitask language understanding","author":"Hendrycks","year":"2020","journal-title":"arXiv preprint"},{"key":"ref31","article-title":"Measuring mathematical problem solving with the math dataset","author":"Hendrycks","year":"2021","journal-title":"arXiv preprint"},{"key":"ref32","article-title":"Lora: Low-rank adaptation of large language models","author":"Edward","year":"2021","journal-title":"arXiv preprint"},{"key":"ref33","article-title":"MiniCPM: Unveiling the potential of small language models with scalable training strategies","author":"Hu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref34","article-title":"Mixtral of experts","author":"Albert","year":"2024","journal-title":"arXiv preprint"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_15"},{"key":"ref36","volume-title":"know saraswati cot dataset","year":"2023"},{"key":"ref37","article-title":"Llava-onevision: Easy visual task transfer","author":"Li","year":"2024","journal-title":"arXiv preprint"},{"key":"ref38","volume-title":"Numinamath tir","author":"Jia","year":"2024"},{"key":"ref39","volume-title":"Numinamath","author":"Jia","year":"2024"},{"key":"ref40","article-title":"From crowdsourced data to high-quality benchmarks: Arena-hard and benchbuilder pipeline","author":"Li","year":"2024","journal-title":"arXiv preprint"},{"key":"ref41","volume-title":"From live data to high-quality benchmarks: The arena-hard pipeline","author":"Li","year":"2024"},{"key":"ref42","volume-title":"Hashimoto. Alpacaeval: An automatic evaluator of instruction-following models","author":"Li","year":"2023"},{"key":"ref43","volume-title":"Slimorca: An open dataset of gpt-4 augmented flan reasoning traces, with verification","author":"Lian","year":"2023"},{"key":"ref44","article-title":"Sphinx-x: Scaling data and parameters for a family of multi-modal large language models","author":"Liu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.624"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4235-6"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72658-3_13"},{"key":"ref48","article-title":"DeepSeek-VL: Towards real-world vision-language understanding","author":"Lu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0182"},{"key":"ref50","article-title":"MathVista: Evaluating mathematical reasoning of foundation models in visual contexts","author":"Lu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref51","volume-title":"Ovis: Structural embedding alignment for multimodal large language model","author":"Lu","year":"2024"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.00392"},{"key":"ref53","article-title":"Small language models: Survey, measurements, and insights","author":"Lu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.02324"},{"key":"ref55","article-title":"Wizardcoder: Empowering code large language models with evol-instruct","author":"Luo","year":"2023","journal-title":"arXiv preprint"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-acl.177"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/WACV48630.2021.00225"},{"key":"ref58","article-title":"Hello GPT-4o","year":"2024","journal-title":"OpenAI"},{"key":"ref59","article-title":"Direct preference optimization: Your language model is secretly a reward model","volume":"36","author":"Rafailov","year":"2024","journal-title":"NeurIPS"},{"key":"ref60","volume-title":"No robots","author":"Rajani","year":"2023"},{"key":"ref61","article-title":"Gpqa: A graduate-level googleproof q&a benchmark","author":"Rein","year":"2023","journal-title":"arXiv preprint"},{"key":"ref62","article-title":"Gpqa: A graduate-level googleproof q&a benchmark","volume-title":"First Conference on Language Modeling","author":"Rein","year":"2024"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00851"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1145\/3694715.3695964"},{"key":"ref65","article-title":"Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context","author":"Team","year":"2024","journal-title":"arXiv preprint"},{"key":"ref66","volume-title":"Internlm: A multilingual language model with progressively enhanced capabilities","year":"2023"},{"key":"ref67","article-title":"Internlm: A multilingual language model with progressively enhanced capabilities","year":"2023","journal-title":"InternLM Team."},{"key":"ref68","volume-title":"Kimi-VL technical report","author":"Team","year":"2025"},{"key":"ref69","article-title":"Cambrian 1: A fully open, vision-centric exploration of multimodal llms","author":"Tong","year":"2024","journal-title":"arXiv preprint"},{"key":"ref70","article-title":"Qwen2-vl: Enhancing vision-language model\u2019s perception of the world at any resolution","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref71","article-title":"CogVLM: Visual expert for pretrained language models","author":"Wang","year":"2023","journal-title":"arXiv preprint"},{"key":"ref72","article-title":"Enhancing the reasoning ability of multimodal large language models via mixed preference optimization","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref73","article-title":"How do your code 11 ms perform? empowering code instruction tuning with high-quality data","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref74","article-title":"Mmlu-pro: A more robust and challenging multi-task language understanding benchmark","author":"Wang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref75","article-title":"Deepseek-vl2: Mixture-ofexperts vision-language models for advanced multimodal understanding","author":"Wu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref76","article-title":"Wizardlm: Empowering large pre-trained language models to follow complex instructions","volume-title":"The Twelfth International Conference on Learning Representations","author":"Xu","year":"2024"},{"key":"ref77","article-title":"Magpie: Alignment data synthesis from scratch by prompting aligned 11 ms with nothing","author":"Xu","year":"2024","journal-title":"arXiv preprint"},{"key":"ref78","article-title":"Powerinfer-2: Fast large language model inference on a smartphone","author":"Xue","year":"2024","journal-title":"arXiv preprint"},{"key":"ref79","volume-title":"Qwen2 technical report","author":"Yang","year":"2024"},{"key":"ref80","article-title":"Qwen2. 5 technical report","author":"Yang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref81","volume-title":"Firefly: A chinese conversational large language model","author":"Yang","year":"2023"},{"key":"ref82","article-title":"Minicpm-v: A gpt-4v level mllm on your phone","author":"Yao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref83","article-title":"Metamath: Bootstrap your own mathematical questions for large language models","author":"Yu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref84","article-title":"Mm-vet: Evaluating large multimodal models for integrated capabilities","author":"Yu","year":"2023","journal-title":"arXiv preprint"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00913"},{"key":"ref86","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01100"},{"key":"ref87","doi-asserted-by":"publisher","DOI":"10.1145\/3627673.3679122"},{"key":"ref88","article-title":"Wings: Learning multimodal 11 ms without text-only forgetting","author":"Zhang","year":"2024","journal-title":"arXiv preprint"},{"key":"ref89","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.906"},{"key":"ref90","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2020"},{"key":"ref91","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2020"},{"key":"ref92","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.762"},{"key":"ref93","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2400"},{"key":"ref94","doi-asserted-by":"publisher","DOI":"10.1145\/3688863.3689575"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11443341.pdf?arnumber=11443341","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T05:15:37Z","timestamp":1777612537000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11443341\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":94,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.00400","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}