How to use monocular image as input using Nuscene? (only use a single view image)

[Rango-T10000]

According to your paper:" With the monocular image or multi-camera images as the
input, the multi-scale features are first extracted by the image encoder, and then lifted to 3D feature volume, ..."

And the demo you put in the github pape shows that OccFormer can use a single view image using SemanticKITTI dataset, my question is that can I use a single view image as input using Nusence dataset? How to modify the code?
I have tried to change the data config in the config file:

And I check the input shape, it is a single image input:

After using one view image ([CAM_FRONT]) as input for inference, I can get the result .pkl files.
But the visualize code can't work. Can you provide me more advice?

[Rango-T10000]

@zhangyp15 Could you please answer my question? Really thanks for your assistance.