
The proposed PromptSync method enhances zero-shot generalization in vision-language models by employing class-aware prototype alignment and discrimination during test-time adaptation.


coremsg

bridging-domain-gaps-in-vision-language-models-through-class-aware-prototype-alignment-and-discrimination


Bridging Domain Gaps in Vision-Language Models through Class-Aware Prototype Alignment and Discrimination


title_rewrite


A generalized linear probe (LP++) that efficiently adapts vision-language models like CLIP to few-shot classification tasks, outperforming recent prompt learning and adapter-based methods while being computationally efficient.


efficient-few-shot-adaptation-of-vision-language-models-using-a-surprisingly-strong-linear-probe-lp-


Efficient Few-Shot Adaptation of Vision-Language Models Using a Surprisingly Strong Linear Probe (LP++)
