TY  - RPRT
TI  - Robustifying Vision Transformer without Retraining from Scratch by Test-Time Class-Conditional Feature Alignment
AU  - Takeshi Kojima
AU  - Yutaka Matsuo
AU  - Yusuke Iwasawa
PY  - 2022
UR  - https://arxiv.org/abs/2206.13951
ID  - 2206.13951
ER  -