1 paper · 2 filters
Geewook Kim, Hodong Lee, Daehee Kim +7
Recent advances in Large Language Models (LLMs) have stimulated a surge of research aimed at extending their applications to the visual domain. While these models exhibit promise i…