Skip to content

Latest commit

 

History

History
32 lines (23 loc) · 764 Bytes

README.md

File metadata and controls

32 lines (23 loc) · 764 Bytes

dwconv

dwconv implements the octree-based depth-wise convolution with CUDA. It speed up the original PyTorch implementation from ocnn by 2.5 times.

The code has been tested on Ubuntu 20.04 with CUDA 11.2 and PyTorch 12.1. After install the required packages, run the following command to install dwconv.

git clone https://github.com/octree-nn/dwconv.git
pip install ./dwconv