search

Pointcept / PointTransformerV3

[CVPR'24 Oral] Official repository of Point Transformer V3 (PTv3)
MIT License
583 stars 30 forks source link

请问作者如何将形状为（b，n，c）的张量预处理成为您要求的字典 #55

Open lsy-study opened 1 month ago

lsy-study commented 1 month ago

""" A data_dict is a dictionary containing properties of a batched point cloud. It should contain the following properties for PTv3:

  1. "feat": feature of point cloud
  2. "grid_coord": discrete coordinate after grid sampling (voxelization) or "coord" + "grid_size"
  3. "offset" or "batch": https://github.com/Pointcept/Pointcept?tab=readme-ov-file#offset """ 特别是"coord" + "grid_size"的获取该如何获取？

另外这边您要求的输入为6维，我是否能将他改变成3维，毕竟代表了x，y，z三个坐标？