AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification

Chun-Fu (Richard) Chen, Quanfu Fan, Rameswar Panda

ICCV 2021 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 2,350 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
use cross-attention fusion instead of no fusion (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.2 -> 81.0Improved
use cross-attention fusion instead of all-attention (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.0 -> 81.0Improved
use cross-attention fusion instead of class token fusion (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.3 -> 81.0Improved
use cross-attention fusion instead of pairwise fusion (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.3 -> 81.0Improved
use CLS tokens for cross-attention instead of averaged patch tokens (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.0 -> 81.0Improved
use patch sizes (12,16) instead of (8,16) (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 80.8 -> 81.0Improved
widen the small branch to 384 dims (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 81.0 -> 80.1Hurt
deepen the small branch (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 81.0 -> 80.7Hurt
stack two cross-attention modules (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 81.0 -> 81.0No clear effect
use more multi-scale encoders with fewer small-branch encoders (ablation)CrossViT-S · ImageNet-1k val (CrossViT-S ablation)top-1 accuracy (%): 81.0 -> 80.9Hurt
add T2T module to replace linear patch embedding (ablation)CrossViT-18 · ImageNet-1k valtop-1 accuracy (%): 82.5 -> 83.0Improved
use convolutional patch embedding (ablation)CrossViT-9 · ImageNet-1k valtop-1 accuracy (%): 73.9 -> 77.1Improved

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