Description: Add support into GuideLLM to benchmark VL Models (Images in, Text out). **Acceptance Critieria**