# Vision-TTT: Efficient and Expressive Visual Representation Learning with Test-Time Training

Source: [Vision-TTT: Efficient and Expressive Visual Representation Learning with Test-Time Training](<https://www.sota2.com/research/paper/69a66bc82d603e6de3e3cb62>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-24T05:00:03.763Z

Authors: Quan Kong, Yanru Xiao, Yuhao Shen, Cong Wang

Published: 2026-02-28T07:31:43.000Z

## Abstract

Learning efficient and expressive visual representation has long been the pursuit of computer vision research. While Vision Transformers (ViTs) gradually replace traditional Convolutional Neural Networks (CNNs) as more scalable vision learners, their applications are plagued by the quadratic complexity of the self-attention mechanism. To address the challenge, we introduce a new linear-time sequence modeling method Test-Time Training (TTT) into vision and propose Vision-TTT, which treats visual sequences as datasets and compresses the visual token sequences in a novel self-supervised learning manner. By incorporating the dual-dataset strategy and Conv2d-based dataset preprocessing, Vision-TTT effectively extends vanilla TTT to model 2D visual correlations with global receptive fields. Extensive experiments show that \\texttt{Vittt-T/S/B} achieve $77.7\\%,81.8\\%,82.7\\%$ Top-1 accuracy on ImageNet classification and also greatly outperform their counterparts on downstream tasks. At $1280\\times1280$ resolution, \\texttt{Vittt-T} reduces FLOPs by $79.4\\%$ and runs $4.72\\times$ faster with $88.9\\%$ less memory than DeiT-T. These results demonstrate the expressiveness and efficiency of Vision-TTT as a strong candidate for the next-generation generic visual backbone.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Instance Segmentation on COCO 2017 (val)](<https://www.sota2.com/research/sota/instance-segmentation-on-coco-2017-val.md>): Instance Segmentation; COCO 2017 (val); APm: 0.428.

- [Semantic Segmentation on ADE20K (val)](<https://www.sota2.com/research/sota/semantic-segmentation-on-ade20k-val.md>): Semantic segmentation; ADE20K (val); mIoU: 49.4.

- [Object Detection on COCO 2017 (val)](<https://www.sota2.com/research/sota/object-detection-on-coco-2017-val.md>): Object Detection; COCO 2017 (val); AP: not recorded.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.00518>)

[PDF](<https://arxiv.org/pdf/2603.00518.pdf>)
