This document provides step-by-step instructions for setting up the environments for Inpaint360GS.
To ensure stability and avoid dependency conflicts between 3D reconstruction and 2D inpainting, we use two isolated Conda environments.
This environment handles 3D Gaussian Splatting (3DGS), Segment-and-Track-Anything, and the core editing pipeline. The following configuration and installation steps have been verified on CUDA 11.8, RTX4090 GPU.
conda create -n inpaint360gs python=3.10 -y
conda activate inpaint360gs
# Install the appropriate version of PyTorch based on your CUDA version.
pip install torch==2.0.0+cu118 torchvision==0.15.1+cu118 torchaudio==2.0.1+cu118 --index-url https://download.pytorch.org/whl/cu118
# Install project dependencies
bash install.shDownload Segmentation Model Weights CropFormer_hornet_3x_03823a.pth， put it under seg/weight/
cd LaMa
conda env create -f conda_env.yml
conda activate lama
conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch -y
pip install pytorch-lightning==1.2.9Download the big-lama.zip weights, extract them under ./LaMa