The number of the point cloud(.ply) file is double of the number of the .label file

[MonaHodaei]

I am testing the semantic3D dataset with pre-trained model of yours, and I am getting .label files as output. However, the number of rows in the .ply file (inside the original_ply folder) is double the number of .label file rows. I think this result is not correct. How can I fix this? Is there any part that I should do re-projection?