Guiding Vision-Language Model Selection for Visual Question-Answering Across Tasks, Domains, and Knowledge Types | Read Paper on Bytez