A focused, lightweight vision model that specializes in generating captions for images. It operates with a narrow but clear purpose — take an image, describe it in text — without trying to do much else. Its small footprint makes it practical for environments where resources are constrained.