I modify the origin Octree-GS SIBR viewer cmakefile to make it possible to install and run it on ubuntu, if you want to install it in Windows, please refer to the origin repository.
- NVIDIA RTX 4070Ti 12G
- Ubuntu 20.04
- CUDA 11.8 (!!! Test in CUDA 11.8 & 11.6, Failed in 11.7)
- Libtorch-2.0-CUDA11.8
- Python 3.9 in Anaconda virtual environment
-
git clone the repository
-
Download the Libtorch with the right CUDA Version(The CUDA Version of Libtorch must be same with the CUDA used in your local environments), and unzip it.
-
Follow the below command
# Dependencies sudo apt install -y libglew-dev libassimp-dev libboost-all-dev libgtk-3-dev libopencv-dev libglfw3-dev libavdevice-dev libavcodec-dev libeigen3-dev libxxf86vm-dev libembree-dev # Project setup cd SIBR_viewers cmake -Bbuild . -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH="the-path-placed-libtorch/libtorch" # add -G Ninja to build faster cmake --build build -j24 --target install
-
After installation, use the GUI just like the origin OctreeGS Readme
This repository is a modified version based on SIBR viewers in Octree-GS. Thanks to their great work!