AI Paper Atlas

Research map / Detection and segmentation / CNN features

EATFormer: Improving Vision Transformer Inspired by Evolutionary Algorithm

Jiangning Zhang, Xiangtai Li, Yabiao Wang, Chengjie Wang, Yibo Yang, Yong Liu, Dacheng Tao

IJCV 2024 · Detection and segmentation › CNN features · approach: Plain and high-resolution ViTs · 59 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 EATFormer-Mobile instead of MobileNetV3-Small 0.75xmobile ViT · ImageNet-1K val (224x224, from scratch)top-1 accuracy: 65.4 -> 69.4Improved
raise input resolution from 224 to 384 for TinyEATFormer-Tiny · ImageNet-1K valtop-1 accuracy: 78.4 -> 80.1Improved
raise input resolution from 224 to 384 for SmallEATFormer-Small · ImageNet-1K valtop-1 accuracy: 83.1 -> 84.3Improved
raise input resolution from 224 to 384 for BaseEATFormer-Base · ImageNet-1K valtop-1 accuracy: 83.9 -> 84.9Improved
use improved EATFormer pyramid design over previous conference columnar EAT (Tiny)EATFormer-Tiny · ImageNet-1K valtop-1 accuracy: +5.7Improved
use improved EATFormer pyramid design over previous conference columnar EAT (Small)EATFormer-Small · ImageNet-1K valtop-1 accuracy: +2.7Improved
use improved EATFormer pyramid design over previous conference columnar EAT (Base)EATFormer-Base · ImageNet-1K valtop-1 accuracy: +1.9Improved
use EATFormer-Base instead of MPViT-Bbackbone · ImageNet-1K valGPU throughput (images/s): 181 -> 329Improved
use EATFormer-Tiny instead of PVT-Tiny backboneMask R-CNN · COCO (1x/3x)box AP: +5.6/+5.6Improved
use EATFormer-Tiny instead of MPViT-T backboneMask R-CNN · COCO (3x schedule)box AP: 44.8 -> 45.4Improved
use EATFormer-Small instead of MPViT-S backboneMask R-CNN · COCO (1x schedule)box AP: 46.4 -> 46.1Hurt
use EATFormer-Tiny instead of VAN-Tiny backboneUperNet · ADE20K valmIoU: 41.1 -> 44.5Improved

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