AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

EA-Vit: Efficient Adaptation for Elastic Vision Transformer

Chen Zhu, Wangbo Zhao, Huiwen Zhang, Yuhao Zhou, Weidong Tang, Shuo Wang, Zhihang Yuan, Yuzhang Shang, Xiaojiang Peng, Kai Wang, Dawei Yang

ICCV 2025 · Architectures and efficiency › Attention and transformers · approach: Hybrid and interpretable ViT designs · 4 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 attention-head elasticity to MLP-only elasticity (ablation)EA-ViT · CIFAR-100minimum MACs (G): 7.80 -> 6.04Improved
add attention-head elasticity to MLP-only elasticity (ablation)EA-ViT · CIFAR-100AUC: 0.505 -> 0.646Improved
add embedding dimension elasticity (ablation)EA-ViT · CIFAR-100minimum MACs (G): 6.04 -> 5.11Improved
add depth elasticity (ablation)EA-ViT · CIFAR-100minimum MACs (G): 5.11 -> 4.60Improved
add depth elasticity (ablation)EA-ViT · CIFAR-100max Top-1 accuracy: 93.31 -> 93.32No clear effect
add embedding elasticity (ablation)EA-ViT · CIFAR-100AUC: 0.646 -> 0.651Improved
add all four elastic dimensions versus MLP only (ablation)EA-ViT · CIFAR-100max Top-1 accuracy: 93.01 -> 93.32Improved
use router-selected submodels instead of manually selected fixed architectures (ablation)EA-ViT · CIFAR-100MACs-accuracy curveImproved
add Curriculum Elastic Adaptation (ablation)EA-ViT · CIFAR-100MACs-accuracy curveMixed
add Curriculum Elastic Adaptation after Stage 2 refinement (ablation)EA-ViT · CIFAR-100MACs-accuracy curveImproved
add Pareto-optimal initialization of router instead of random (ablation)EA-ViT · CIFAR-100MACs-accuracy curveImproved
use EA-ViT instead of best prior method on Flowers-102 at 8 GMACsEA-ViT · Flowers-102Top-1 accuracy: 73.80 -> 85.39Improved

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