tusen-ai / SST

Code for a series of work in LiDAR perception, including SST (CVPR 22), FSD (NeurIPS 22), FSD++ (TPAMI 23), FSDv2, and CTRL (ICCV 23, oral).
Apache License 2.0
801 stars 102 forks source link

The question about voxelize #64

Closed Fengshihao1 closed 2 years ago

Fengshihao1 commented 2 years ago

作者,您好! 非常感谢您的工作! 我对体素化的过程有一点困惑。在您的SST工作中,对输入的雷达点会首先进行voxelize: 在mmdet3d/models/detectors/dynamic_voxelnet.py中的: def voxelize(self, points): """Apply dynamic voxelization to points. Args: points (list[torch.Tensor]): Points of each sample. Returns: tuple[torch.Tensor]: Concatenated points and coordinates. """ 这里输入的points是列表的形式,请问列表的长度是batchsize的大小吗?

Abyssaledge commented 2 years ago

是的.