RST: Rough set Transformer for 3D Point Cloud Learning

This is a Pytorch implementation of RST

Requirements

python >= 3.6

pytorch >= 1.6

h5py

scikit-learn

and

pip install pointnet2_ops_lib/.
python install setup.py

Modified from PCT https://github.com/Strawberry-Eat-Mango/PCT_Pytorch.

Models

We provide two models to train, our model is placed in RSTmodel.py

And If you want to reproduce PCT, please choose PCTmodel.py

Example training and testing

# train
python main.py --exp_name=train --num_points=1024 --use_sgd=True --batch_size 32 --epochs 600 --lr 0.0001

# test
python main.py --exp_name=test --num_points=1024 --use_sgd=True --eval=True --model_path=checkpoints/train/models/model.t7 --test_batch_size 8