1 paper
Nilakshan Kunananthaseelan, Jing Zhang, Mehrtash Harandi
We introduce a language-grounded visual prompting method to adapt the visual encoder of vision-language models for downstream tasks. By capitalizing on language integration, we dev…