Text this: LLaVA-docent: Instruction tuning with multimodal large language model to support art appreciation education