AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Fcaformer: Forward Cross Attention in Hybrid Vision Transformer

Haokui Zhang, Wenze Hu, Xiaoyu Wang

ICCV 2023 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 10 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 naive forward cross attention without scale factors (ablation)ConvViT (hybrid, global attention) · ImageNet-1K valtop-1 accuracy (%): 79.4 -> 79.4No clear effect
add learnable scale factors (LSFs) for cross tokens (ablation)FcaFormer · ImageNet-1K valtop-1 accuracy (%): 79.4 -> 79.9Improved
add token merge and enhancement (TME) module (ablation)FcaFormer · ImageNet-1K valtop-1 accuracy (%): 79.9 -> 80.3Improved
add early convolution stages (ablation)Swin-1G · ImageNet-1K valtop-1 accuracy (%): 78.4 -> 78.8Improved
replace window attention with global attention (ablation)ConvSwin · ImageNet-1K valtop-1 accuracy (%) and latency: 78.8 -> 79.4; latency 269 -> 300 msMixed
replace DeiT-T with forward cross attention (FcaFormer-L0)DeiT-Tiny · ImageNet-1K valtop-1 accuracy (%): 72.2 -> 74.3Improved
use FcaFormer-L2 instead of ConvNext-Tinyhybrid backbone · ImageNet-1K valtop-1 accuracy (%): 82.1 -> 83.1Improved
use FcaFormer-L2 backbone in UperNet instead of ConvNext-TUperNet · ADE20KmIoU: 46.7 -> 47.6Improved
use FcaFormer-L2 backbone in Mask-RCNN instead of ConvNext-TMask-RCNN 3x · COCO 2017box AP: 46.2 -> 47.0Improved
add learnable scale factors to calibrate cross tokens (attention analysis) (ablation)FcaFormer · attention score mapsattention to cross tokensImproved

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