How to save the feature representations of all the images and pointcloud pairs?

[shahriar0651]

What is the feature?

I want to utilize the pre-trained weights of the feature extractor and save the feature representation pairs from both the images and point clouds before fusing them. Could anyone help me by giving me directions?

Any other context?

No response

[Petros626]

https://mmdetection3d.readthedocs.io/en/latest/user_guides/visualization.html