Composable Visual Tokenizers with Generator-Free Diagnostics of Learnability

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Zhao, Bingchen, Guo, Qiushan, Wang, Ye, Huang, Yixuan, Zhai, Zhonghua, Tian, Yu
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2602.03339