Grounding Visual Representations with Texts for Domain Generalization | Read Paper on Bytez