how to create the pickle file for a custom dataset?

[gynaa]

Hello everyone,

I want to train some models on a custom dataset. So I have to create a pickle file with the data. I looked at the kitti pickle file and it has a few parameters, that I am not sure where to get them from: num_points_in_gt, gt_idx and group_id.

Could someone help me out here please?

[JingweiZhang12]

num_points_in_gt, gt_idx and group_id. is the annotation of specific datasets. If you don't have these annotations, you can ignore it.