Please follow the 3D-GS to install the relative packages.

conda create -n 4dedit python=3.8 
conda activate 4dedit

pip install -r requirements.txt
pip install -e submodules/depth-diff-gaussian-rasterization
pip install -e submodules/simple-knn

In our environment, we use pytorch=1.13.1+cu116.

Follow the instructions at this link to create the environment and install dependencies. Only follow the commands up to tinycudann. After the dependencies have been installed, return here.

Once you have finished installing dependencies, including those for gsplat, you can install Instruct-GS2GS using the following command:

pip install git+https://github.com/cvachha/instruct-gs2gs

For synthetic scenes: The dataset provided in D-NeRF is used. You can download the dataset from dropbox.

For real dynamic scenes: The dataset provided in HyperNeRF is used. You can download scenes from Hypernerf Dataset and organize them as Nerfies.

For training hypernerf scenes such as virg/broom: Pregenerated point clouds by COLMAP are provided here. Just download them and put them in to correspond folder.

├── data
│   | dnerf 
│     ├── mutant
│     ├── standup 
│     ├── ...
│   | hypernerf
│     ├── interp
│     ├── virg
│   

For training synthetic scenes such as trex, run

python train.py -s data/dnerf/trex --port 6017 --expname "dnerf/trex" --configs arguments/dnerf/trex.py --prompt "Turn the skeleton into gold and the stone into red" --dataset_change_iter 10000

Run the following script to render the images.

python render.py --model_path "output/dnerf/trex/"  --skip_train --configs arguments/dnerf/trex.py  &

Thank you to the authors of the original projects for providing the codebase. "InstructGS2GS" and "4D Gaussian Splatting for Real-Time Dynamic Scene Rendering"