InstructBLIP Cog package InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning This is an implemention of a Cog package for the InstructBLIP image-to-text model. Demo