TY  - RPRT
TI  - CLIP Itself is a Strong Fine-tuner: Achieving 85.7% and 88.0% Top-1 Accuracy with ViT-B and ViT-L on ImageNet
AU  - Xiaoyi Dong
AU  - Jianmin Bao
AU  - Ting Zhang
AU  - Dongdong Chen
AU  - Shuyang Gu
AU  - Weiming Zhang
AU  - Lu Yuan
AU  - Dong Chen
AU  - Fang Wen
AU  - Nenghai Yu
PY  - 2022
UR  - https://arxiv.org/abs/2212.06138
ID  - 2212.06138
ER  -