StableLLaVA: Enhanced Visual Instruction Tuning with Synthesized Image-Dialogue Data

Open in new window