AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Token-Label Alignment for Vision Transformers

Han Xiao, Wenzhao Zheng, Zheng Zhu, Jie Zhou, Jiwen Lu

ICCV 2023 · Architectures and efficiency › Attention and transformers · approach: Token merging and masked pretraining · 7 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 token-label alignment to CutMix targetsDeiT-S · ImageNet-1k valtop-1 accuracy (%): 79.8 -> 80.6Improved
align labels layer by layer instead of only last-layer attention (ablation)DeiT-S · ImageNet-1k valtop-1 accuracy (%): 80.1 -> 80.6Improved
align only at some middle layers (ablation)DeiT-S · ImageNet-1k valtop-1 accuracy (%): 79.8 -> 80.2Improved
keep normalization in alignment (ablation)TL-Align · ImageNet-1k valtop-1 accuracy (%): 80.3 -> 80.6Improved
apply TL-Align to MixUp+CutMix (ablation)DeiT-S · ImageNet-1k valtop-1 accuracy (%): 79.8 -> 80.2Improved
apply TL-Align to block-wise mixing (ablation)DeiT-S · ImageNet-1k valtop-1 accuracy (%): 80.0 -> 80.3Improved
use TL-Align instead of TransMix on same recipeDeiT-S · ImageNet-1k valtop-1 accuracy (%): 80.1 -> 80.6Improved
use saliency or attention-based mixing (SaliencyMix, Attentive-CutMix) instead of CutMixDeiT-S · ImageNet-1k valtop-1 accuracy (%): 79.8 -> 79.2 (SaliencyMix), 77.5 (Attentive-CutMix)Hurt
add TL-Align on Swin-T for COCO detectionSwin-T Cascade Mask R-CNN · COCO valAP box: 50.4 -> 50.5No clear effect
add TL-Align to DeiT-TDeiT-T · ImageNet-Atop-1 accuracy (%): 7.7 -> 6.1Hurt
add TL-Align to DeiT-BDeiT-B · ImageNet-Rtop-1 accuracy (%): 45.3 -> 44.4Hurt
add TL-Align training costDeiT-S · training speedimage/s: 322 -> 311Hurt

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