Advancing Cross-domain Discriminability in Continual Learning of Vision-Language Models | Read Paper on Bytez