1 paper
Jiaying Shi, Xuetong Xue, Shenghui Xu
The recent CLIP-based methods have shown promising zero-shot and few-shot performance on image classification tasks. Existing approaches such as CoOp and Tip-Adapter only focus on…