ma-xu/pointMLP-pytorch

Use without CUDA?

Closed this issue · 2 comments

Hello!
Would it be possible to use your code without the need of CUDA?

ma-xu commented

@simon-at-hypothetic
Sure, You can siply call the defined farthest_point_sample function instead of furthest_point_sample in pointnet2_ops lib.
Here is a simple example (tested, works but will be slower):
https://web.northeastern.edu/smilelab/xuma/pointMLP/pointmlp_nocuda.py

Please let me know if you have any further questions.

thats great, the code you shared worked for me, thanks!