A PyTorch implemetation of PointCNN (CS294-131 Course Proj)
Switch branches/tags
Nothing to show
Clone or download
Latest commit 6ec6c29 Apr 27, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
doc modified Apr 16, 2018
models modified Apr 16, 2018
part_seg modified Apr 16, 2018
sem_seg modified Apr 16, 2018
utils debug Apr 19, 2018
.gitignore Initial commit Mar 19, 2018
LICENSE Initial commit Mar 19, 2018
README.md Update README.md Apr 27, 2018
evaluate.py modified Apr 16, 2018
provider.py modified Apr 16, 2018
train.py modified Apr 16, 2018
train_pytorch.py modified Apr 23, 2018



This is a PyTorch implementation of PointCNN. It is as efficent as the origin Tensorflow implemetation and achieves same accuracy on both classification and segmentaion jobs. See the following references for more information:

Yangyan Li, Rui Bu, Mingchao Sun, Baoquan Chen
arXiv preprint arXiv:1801.07791, 2018.



We've tested code on ModelNet40 only.

python train_pytorch.py


Our code is released under MIT License (see LICENSE file for details).