前言
目前，R2-Gaussian相关代码已发布，记录了在Windows环境下配置遇到的问题
1. Installation
We recommend using Conda to set up an environment. We tested the code on Win10 with an RTX 4080 GPU. For installation issues on other platforms, please refer to Gaussian Splatting.
# Download code
git clone https://github.com/Ruyi-Zha/r2_gaussian.git --recursive
# Install environment
SET DISTUTILS_USE_SDK=1 # Windows only
# change to your env path
d:
cd d:\anaconda\env
conda env create --file environment.yml
conda activate r2_gaussian
# Install TIGRE for data generation and initialization
# download the TIGRE package
https://github.com/CERN/TIGRE/archive/refs/tags/v2.3.zip
unzip v2.3.zip
# move the TIGRE fold to your conda path (eg. d:\anaconda\env)
pip install TIGRE-2.3/Python --no-build-isolation
install the xray-gaussian-rasterization-voxelization
# move to your g2_gaussian project path
# "M_PI" is undefined would occur when compile the package in Win10
# Add # define M-PI 3.14159265358979323846 in the backward.cu and forwards.cu files
# compile the package
pip install r2_gaussian\submodules\xray-gaussian-rasterization-voxelization
