This repository is the implementation of "Adaptive GMM Convolution for Point Cloud Learning"

cd utils
sh complile_op.sh

This repository is partially dependent on 'pytorch', 'torch_geometric' and 'torch_points3d'.

Please see 'trainval.y' for classification and segmentation tasks, and 'trainval_normal.py' for normal estimation task.
Please see 'configure.py' for the settings.
And see the 'dataset' directories for the data processing.

Part of the codes refers to the KPConv and RandLA-Net

This repository will continue updating.