GraphVL: Graph-Enhanced Semantic Modeling via Vision-Language Models for Generalized Class Discovery

VLM
Re-assign community