AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Locality Guidance for Improving Vision Transformers on Tiny Datasets

Kehan Li, Runyi Yu, Zhennan Wang, Li Yuan, Guoli Song, Jie Chen

ECCV 2022 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 63 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 locality guidance from a lightweight CNNDeiT-Ti · CIFAR-100top-1 accuracy: 65.08 -> 78.15Improved
add locality guidance from a lightweight CNNDeiT-Ti · Oxford Flowerstop-1 accuracy: 50.06 -> 68.50Improved
add locality guidance from a lightweight CNNPVTv2-B0 · CIFAR-100top-1 accuracy: 77.44 -> 79.30Improved
add locality guidance from a lightweight CNNPVT-Ti · Chaoyangtop-1 accuracy: 73.68 -> 85.65Improved
add locality guidance (ours) instead of L_drlocT2T-ViT-14 · CIFAR-100 (100 epochs)top-1 accuracy: 68.03 -> 77.84Improved
use feature guidance instead of logit distillation (DeiT-Hard)DeiT-Tiny · CIFAR-100top-1 accuracy: 73.25 -> 78.15Improved
raise guidance ratio R from 0.25 to 1.0 (use all CNN features) (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 70.56 -> 77.29Improved
add guidance loss weight beta (0.0 -> 2.5) (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 65.08 -> 77.29Improved
use learnable linear projection instead of SP (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 67.36 -> 77.29Improved
use learnable linear projection instead of AT (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 73.51 -> 77.29Improved
use a stronger guidance CNN, ResNet-110 instead of ResNet-56 (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 77.29 -> 76.62No clear effect
use a weaker guidance CNN, ResNet-20 instead of ResNet-56 (ablation)DeiT-Tiny + guidance · CIFAR-100 (100 epochs)top-1 accuracy: 77.29 -> 72.91Hurt

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