AI Paper Atlas

Research map / Detection and segmentation / CNN features

ViT$^3$: Unlocking Test-Time Training in Vision

Dongchen Han, Yining Li, Tianyu Li, Zixuan Cao, Ziming Wang, Jun Song, Yu Cheng, Bo Zheng, Gao Huang

CVPR 2026 · Detection and segmentation › CNN features · approach: Convolution-transformer backbones · 14 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 MSE instead of dot-product inner loss (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 79.2No clear effect
replace inner loss with MAE (L1) (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 76.5Hurt
use sequential mini-batch inner updates instead of full batch (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 78.1Hurt
increase inner epochs from 1 to 3 (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 79.2; FPS 1315 -> 787Mixed
inner learning rate 1.0 versus 0.1 (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 77.5 -> 78.9Improved
very large inner learning rate (5.0, 10.0) (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 76.7*Hurt
token-wise dynamic inner learning rate (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 78.7Hurt
widen inner MLP hidden ratio from 1 to 4 (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 79.6Improved
deepen inner MLP from 2 to 3 layers (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 77.5Hurt
remove output layer of inner MLP, SiLU(FC(x)) (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 79.4Improved
add residual connection to inner MLP (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 78.8No clear effect
use depthwise conv inner model instead of 2-layer MLP (ablation)DeiT-S TTT · ImageNet-1K valtop-1 accuracy: 78.9 -> 80.1Improved

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