search

mit-han-lab / pvcnn

[NeurIPS 2019, Spotlight] Point-Voxel CNN for Efficient 3D Deep Learning
https://pvcnn.mit.edu/
MIT License
644 stars 130 forks source link

Question about voxelization modul (compares to spvoxelize in torchsparse) #81

Closed sIHURs closed 3 months ago

sIHURs commented 3 months ago

Hi~

I'm just starting to learn some deep learning concepts for point clouds. I have a project that requires voxelization as an operation, and I need to incorporate voxelization as a layer in a PyTorch model.

Due to some environment issues, the modules from PVCNN have problems with import (even after running for a long time, the import is not successful). However, I successfully downloaded TorchSparse. I noticed that TorchSparse has a method torchsparse.nn.functional.spvoxelize that seems very similar to the voxelization module in PVCNN. The TorchSparse wiki does not provide information about the counts parameter, I'm not sure what it is.

The question is, if I have coords, feats, and resolution, how can I use spvoxelize to get a voxelized output with the shape (b, c, resolution, resolution, resolution)?

thanks very much yifan

zhijian-liu commented 3 months ago

@sIHURs, this repository is no longer actively maintained. Thanks for your interest in this work.