{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/ubimoe-a-ubiquitous-mixture-of-experts-vision","title":"UbiMoE: A Ubiquitous Mixture-of-Experts Vision Transformer Accelerator With Hybrid Computation Pattern on FPGA","arxiv_id":"2502.05602","date":"2025-02-08","proceeding":null,"authors":["Jiale Dong","Wenqi Lou","Zhendong Zheng","Yunji Qin","Lei Gong","Chao Wang","Xuehai Zhou"],"abstract":"Compared to traditional Vision Transformers (ViT), Mixture-of-Experts Vision Transformers (MoE-ViT) are introduced to scale model size without a proportional increase in computational complexity, making them a new research focus. Given the high performance and reconfigurability, FPGA-based accelerators for MoE-ViT emerge, delivering substantial gains over general-purpose processors. However, existing accelerators often fall short of fully exploring the design space, leading to suboptimal trade-offs between resource utilization and performance. To overcome this problem, we introduce UbiMoE, a novel end-to-end FPGA accelerator tailored for MoE-ViT. Leveraging the unique computational and memory access patterns of MoE-ViTs, we develop a latency-optimized streaming attention kernel and a resource-efficient reusable linear kernel, effectively balancing performance and resource consumption. To further enhance design efficiency, we propose a two-stage heuristic search algorithm that optimally tunes hardware parameters for various FPGA resource constraints. Compared to state-of-the-art (SOTA) FPGA designs, UbiMoE achieves 1.34x and 3.35x throughput improvements for MoE-ViT on Xilinx ZCU102 and Alveo U280 platforms, respectively, while enhancing energy efficiency by 1.75x and 1.54x. Our implementation is available at https://github.com/DJ000011/UbiMoE.","url_abs":"https://arxiv.org/abs/2502.05602v3","url_pdf":"https://arxiv.org/pdf/2502.05602v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"links_only","authors_date_abstract":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license), from the Kaggle arXiv metadata snapshot of 2026-09-12"},"code_links":[{"paper_slug":"ubimoe-a-ubiquitous-mixture-of-experts-vision","repo_url":"https://github.com/dj000011/ubimoe","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"ubimoe-a-ubiquitous-mixture-of-experts-vision","repo_url":"https://github.com/dj000011/coqmoe","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2502.05602","atlas_url":"https://app.syntology.ai/?focus=2502.05602","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}