研究人员开发了FlexViT，这是一款灵活的基于FPGA的加速器，旨在提高边缘设备上视觉Transformer (ViT) 模型的效率。该加速器通过将现代ViT的异构架构（结合了卷积层和全连接层）映射到一个统一的INT8 GEMM引擎上，解决了其异构架构带来的挑战。FlexViT采用双模式数据流和深度优先分块策略来优化性能并降低内存带宽需求。评估显示，FlexViT在加速器执行的层上可实现高达2.74倍的加速，与纯CPU执行相比，整体加速可达1.40倍。 AI
影响 增强了在资源受限的边缘设备上部署复杂视觉Transformer模型的可行性。
排序理由 该集群描述了一篇详细介绍AI模型新硬件加速器的研究论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2606.31938v1 Announce Type: cross Abstract: Deploying Vision Transformer (ViT) models on edge platforms remains challenging due to their high computational demands and the architectural heterogeneity of modern hybrid ViT models, which incorporate both fully connected and co…
Deploying Vision Transformer (ViT) models on edge platforms remains challenging due to their high computational demands and the architectural heterogeneity of modern hybrid ViT models, which incorporate both fully connected and convolutional layers. This heterogeneity leads to si…