Qwen3-VL image analysis using vLLM on DGX Spark

I'm at the stage with my poster generation project where I am analysing poster images to create training data. It turns out that vLLM is an excellent inference engine for this use case due to its supp