From Images to Textual Prompts: Zero-Shot Visual Question Answering With Frozen Large Language Models | Read Paper on Bytez