提出新评估框架，精准筛选边缘设备上省电的视觉Transformer模型。
Energy-Efficient Vision Transformer Inference for Edge-AI Deployment
- 分两阶段评估：先用NetScore筛选模型，再用SAM实测能效。
- 混合模型LeViT_Conv_192在边缘设备上节能53%，比基线更省电。
- 轻量化蒸馏模型适合移动端GPU，能效比提升显著。
视觉Transformer（ViT）在资源受限设备上的部署日益增多，单纯以精度评估已不足够。本文提出一种两阶段评估框架，结合与设备无关的模型筛选和与设备相关的实测指标。我们在ImageNet-1K和CIFAR-10数据集上对13个ViT模型进行测试，分别在NVIDIA Jetson TX2（边缘设备）和RTX 3050（移动GPU）上运行推理。第一阶段使用NetScore进行模型预选；第二阶段采用可持续准确率指标（SAM）进行排序。结果表明，混合模型LeViT_Conv_192在TX2上相比基线模型能效提升达53%（如在CIFAR-10上SAM5=1.44），而蒸馏模型TinyViT-11M_Distilled在移动GPU上表现优异（如在RTX 3050/CIFAR-10上SAM5=1.72，ImageNet-1K上SAM5=0.76）。
原文摘要 · Abstract (English)
The growing deployment of Vision Transformers (ViTs) on energy-constrained devices requires evaluation methods that go beyond accuracy alone. We present a two-stage pipeline for assessing ViT energy efficiency that combines device-agnostic model selection with device-related measurements. We benchmark 13 ViT models on ImageNet-1K and CIFAR-10, running inference on NVIDIA Jetson TX2 (edge device) and an NVIDIA RTX 3050 (mobile GPU). The device-agnostic stage uses the NetScore metric for screening; the device-related stage ranks models with the Sustainable Accuracy Metric (SAM). Results show that hybrid models such as LeViT_Conv_192 reduce energy by up to 53% on TX2 relative to a ViT baseline (e.g., SAM5=1.44 on TX2/CIFAR-10), while distilled models such as TinyViT-11M_Distilled excel on the mobile GPU (e.g., SAM5=1.72 on RTX 3050/CIFAR-10 and SAM5=0.76 on RTX 3050/ImageNet-1K).
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。