AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

An Experimental Study on Exploring Strong Lightweight Vision Transformers via Masked Image Modeling Pre-training

Jin Gang Gao, Shubo Lin, Shaoru Wang, Yutong Kou, Zeming Li, Liang Li, Congxuan Zhang, Xiaoqin Zhang, Yizheng Wang, Weiming Hu

IJCV 2025 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 13 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-map distillation from MAE ViT-Base on the last layer during MAE pre-training (ablation)MAE ViT-Tiny · ADE20KmIoU: 34.5 -> 42.9Improved
add attention-map distillation from MAE ViT-Base during MAE pre-training (ablation)MAE ViT-Tiny · Petstop-1 accuracy: 76.5 -> 89.1Improved
add attention-map distillation during MAE pre-training (ablation)MAE ViT-Tiny · ImageNet-1K valtop-1 accuracy: 78.0 -> 78.3Improved
decouple distillation from reconstruction by connecting the decoder to layer 8 instead of 12 (D2-MAE) (ablation)D-MAE ViT-Tiny · Petstop-1 accuracy: 89.1 -> 90.4Improved
decouple distillation by connecting the decoder to layer 8 instead of 12 (ablation)D-MAE ViT-Tiny · COCO detectionbox AP: 42.0 -> 42.5Improved
add MAE reconstruction loss to attention distillation (ablation)D2-MAE ViT-Tiny · Petstop-1 accuracy: 84.5 -> 90.4Improved
add MAE reconstruction loss to attention distillation (ablation)D2-MAE ViT-Tiny · COCO detectionbox AP: 39.6 -> 42.5Improved
pre-train on IN21K instead of IN1K (ablation)MAE ViT-Tiny · ImageNet-1K valtop-1 accuracy: 78.0 -> 78.0No clear effect
pre-train on IN21K instead of IN1K (ablation)MoCo-v3 ViT-Tiny · ImageNet-1K valtop-1 accuracy: 76.8 -> 76.9No clear effect
reduce pre-training data to 1% of IN1K (ablation)MoCo-v3 ViT-Tiny · ImageNet-1K valtop-1 accuracy: 76.8 -> 76.2Hurt
reduce pre-training data to 1% of IN1K (ablation)MAE ViT-Tiny · ImageNet-1K valtop-1 accuracy: 78.0 -> 77.9No clear effect
use global average pooling instead of class token for fine-tuning (ablation)MoCo-v3 ViT-Tiny · ImageNet-1K valtop-1 accuracy: 73.7 -> 76.8Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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