AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

EdgeViTs: Competing Light-Weight CNNs on Mobile Devices with Vision Transformers

Junting Pan, Adrian Bulat, Fuwen Tan, Xiatian Zhu, Lukasz Dudziak, Hongsheng Li, Georgios Tzimiropoulos, Brais Martinez

ECCV 2022 · Architectures and efficiency › Attention and transformers · approach: Hybrid and interpretable ViT designs · 326 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add local aggregation (with global sparse attention) rather than a single PVT-style attention block (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 72.7 -> 74.4Improved
use depthwise-conv local aggregation instead of locally-grouped self-attention (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 74.0 -> 74.4Improved
use depthwise-conv local aggregation instead of locally-grouped self-attention (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)CPU latency (ms): 36.1ms -> 32.8msImproved
sample delegate tokens by center instead of max (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 74.3 -> 74.4No clear effect
sample delegate tokens by center instead of avg (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)CPU latency (ms): 34.5ms -> 32.8msImproved
sample delegate tokens by avg instead of center (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 74.4 -> 74.5No clear effect
add local propagation (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 73.9 -> 74.4Improved
add local propagation (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)CPU latency (ms): 32.4ms -> 32.8msHurt
use transposed convolution instead of bilinear upsampling for local propagation (ablation)EdgeViT-XXS · ImageNet-1K val (ablation, ~0.5 GFLOPs)top-1 accuracy (%): 74.1 -> 74.4Improved
EdgeViT-XXS instead of PVT-v2-B0 as backboneEdgeViT-XXS · ADE20K val (Semantic FPN)mIoU (%): 37.2 -> 39.7Improved
EdgeViT-S instead of PVT-v2-B1 as backboneEdgeViT-S · ADE20K val (Semantic FPN)mIoU (%): 42.5 -> 45.9Improved
EdgeViT-XXS instead of PVT-v2-B0 as RetinaNet backboneEdgeViT-XXS · COCO val2017 (RetinaNet 1x)AP: 37.2 -> 38.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy