Hi there,

I wonder if you can share an example of running offline batch inference with vLLM on video understanding, where the videos are loaded from local paths?

Thanks you so much!