Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How to obtain voxel labels? #185

Open
Haerxu opened this issue Sep 5, 2023 · 2 comments
Open

How to obtain voxel labels? #185

Haerxu opened this issue Sep 5, 2023 · 2 comments

Comments

@Haerxu
Copy link

Haerxu commented Sep 5, 2023

Hi,

Thanks for releasing code! I have a question about how to obtain voxel labels? In the code, you only compute voxel coordinates and features. How to obtain voxel labels?
image

@thangvubk
Copy link
Owner

Voxel label is not computed. In fact, we convert the voxel output prediction to point-level prediction then compute the loss.

output_feats = output_feats[input_map.long()]

@Haerxu
Copy link
Author

Haerxu commented Nov 2, 2023

Hi,

In the voxelization code softgroup/ops/src/voxelize/voxelize.cpp, there are 5 choices of voxelization and 4 is chosen in the code. Do you know other meaning of the voxxelization? I want to use the voxel labels, but I don't think mode=4 can output voxel label becuas it use average pooling within each voxel.

mode 0=guaranteed unique 1=last item(overwrite) 2=first item(keep) 3=sum, 4=mean
'''

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants