An example of installation is shown below:
git clone https://github.com/KU-CVLAB/3DFuse
cd 3DFuse
conda create -n 3DFuse python=3.8
conda activate 3DFuse
pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113
conda install pytorch3d -c pytorch3d
If PyTorch3D installation fails, please try
pip install git+https://github.com/facebookresearch/pytorch3d@c8af1c45ca9f4fdd4e59b49172ca74983ff3147a#egg=pytorch3d
or follow the instructions here.
And then,
pip install -r requirements.txt
The weights required for the operation of our framework can be downloaded from our HuggingFace page. Please modify the env.json
file with the path of the weight file afterwards.
mkdir weights
cd weights
wget https://huggingface.co/jyseo/3DFuse_weights/resolve/main/models/3DFuse_sparse_depth_injector.ckpt