AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

TinyMIM: An Empirical Study of Distilling MIM Pre-Trained Models

Sucheng Ren, Fangyun Wei, Zheng Zhang, Han Hu

CVPR 2023 · Architectures and efficiency › Pruning and efficient tuning · approach: Efficient ViT adaptation · 47 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
Distill token relations rather than class tokens or features (ablation)TinyMIM ViT-T, ViT-S, and ViT-B · ImageNet-1Ktop-1 accuracy: 75.8 (+4.2); 83.0 (+3.1); 85.0 (+1.6)Improved
Distill token relations rather than class tokens or features (ablation)TinyMIM ViT-B · ADE20KmIoU: 52.2 (+4.1)Improved
Add MAE reconstruction loss to class-token distillation (ablation)TinyMIM ViT-B · ImageNet-1Ktop-1 accuracy: 80.6 -> 82.1Improved
Distill FFN or attention features before, rather than after, the residual connection (ablation)TinyMIM ViT-B · ImageNet-1Ktop-1 accuracy: 84.2 -> 81.8; 84.1 -> 81.3Hurt
Apply softmax to the Q-K and V-V relation targets (ablation)TinyMIM ViT-B · ImageNet-1Ktop-1 accuracy: 84.4 -> 84.6Improved
Use Q-K and V-V relations instead of Q-Q, K-K, and V-V relations (ablation)TinyMIM ViT-B · ImageNet-1Ktop-1 accuracy: 84.4 -> 84.4No clear effect
Select teacher block 18 rather than block 24 (ablation)ViT-B distilled from MAE ViT-L · ImageNet-1Ktop-1 accuracy: 84.4 -> 84.6Improved
Select teacher block 18 rather than block 24 (ablation)ViT-B distilled from MAE ViT-L · ADE20KmIoU: 50.0 -> 52.2Improved
Use MAE ViT-L directly instead of MAE ViT-B as teacher (ablation)ViT-S · ImageNet-1Ktop-1 accuracy: 82.3 -> 82.1Hurt
Use sequential teacher ViT-L -> TinyMIM ViT-B instead of MAE ViT-L directly (ablation)ViT-S · ImageNet-1Ktop-1 accuracy: 82.1 -> 82.6Improved
Use sequential teacher MAE ViT-B -> TinyMIM ViT-S instead of MAE ViT-B directly (ablation)ViT-T · ImageNet-1Ktop-1 accuracy: 74.4 -> 75.0Improved
Use a raw image rather than a masked image as distillation input (ablation)TinyMIM ViT-B · ImageNet-1Ktop-1 accuracy: 83.9 -> 84.6Improved

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