Abstract: Adapter-based parameter-efficient transfer learning has achieved exciting results in vision-language models. Traditional adapter methods often require training or fine-tuning, facing ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results