3D encoder of SPVCNN ?

[sandeepnmenon]

As per the paper, I read that
The 3D encoder is a modified SPVCNN [10] (voxel size 0.1) with fewer parameters, whose hidden dimensions are 64 for SemanticKITTI and 128 for NuScenes to speed up the network.

However the code has

2DPASS/network/arch_2dpass.py

Line 8 in 80b8646

from network.baseline import get_model as SPVCNN

the baseline model in the code.
The change was done 2 months back. Was this intentional?
Which should be run to replicate the SOTA results mentioned in the paper?

[mp0312]

i have the same question