InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning Paper • 2305.06500 • Published May 11, 2023 • 4