CViT通过分块前馈网络和级联注意力，实现高效低耗视觉识别。
CascadedViT: Cascaded Chunk-FeedForward and Cascaded Group Attention Vision Transformer
- 采用分块前馈与级联注意力结构，降低计算开销。
- 在ImageNet-1K上比EfficientViT-M5少15%算力，能耗降3.3%。
- 适合手机、无人机等低功耗设备部署，算力效率领先。
视觉变压器（ViTs）在众多计算机视觉任务中表现优异，但其高计算、内存和能耗限制了在资源受限平台的部署。本文提出轻量级高效的视觉变压器架构Cascaded-ViT（CViT），引入新型前馈网络设计——级联分块前馈网络（CCFFN）。通过分割输入特征，该设计在不损失准确率的前提下提升参数与浮点运算效率。ImageNet-1K实验表明，我们的CViT-XL模型达到75.5% Top-1准确率，相比EfficientViT-M5减少15% FLOPs和3.3%能耗。在多种模型规模下，CViT系列均表现出最低能耗，适用于电池受限设备如手机与无人机。此外，基于新指标准确率-每FLOP（APF）评估，CViT模型始终位列榜首；尤其CViT-L比EfficientViT-M2高2.2%准确率，且具有相当的APF得分。
原文摘要 · Abstract (English)
Vision Transformers (ViTs) have demonstrated remarkable performance across a range of computer vision tasks; however, their high computational, memory, and energy demands hinder deployment on resource-constrained platforms. In this paper, we propose \emph{Cascaded-ViT (CViT)}, a lightweight and compute-efficient vision transformer architecture featuring a novel feedforward network design called \emph{Cascaded-Chunk Feed Forward Network (CCFFN)}. By splitting input features, CCFFN improves parameter and FLOP efficiency without sacrificing accuracy. Experiments on ImageNet-1K show that our \emph{CViT-XL} model achieves 75.5\% Top-1 accuracy while reducing FLOPs by 15\% and energy consumption by 3.3\% compared to EfficientViT-M5. Across various model sizes, the CViT family consistently exhibits the lowest energy consumption, making it suitable for deployment on battery-constrained devices such as mobile phones and drones. Furthermore, when evaluated using a new metric called \emph{Accuracy-Per-FLOP (APF)}, which quantifies compute efficiency relative to accuracy, CViT models consistently achieve top-ranking efficiency. Particularly, CViT-L is 2.2\% more accurate than EfficientViT-M2 while having comparable APF scores.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。