Physic-plausible Part Generation

Package required: accelerate + diffusers + other necessary packages (listed in requirements.txt) + pytorch3d + kaolin

Install pytorch following instructions on pytorch.org

Install pytorch3d

pip install scikit-image matplotlib imageio plotly opencv-python

pip install black usort flake8 flake8-bugbear flake8-comprehensions

We recomment install pytorch3d from source pip install "git+https://github.com/facebookresearch/pytorch3d.git"

Install kaolin (adjust the command based on the pytorch and cuda version)

pip install kaolin==0.15.0 -f https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.0.1_cu117.html

Install other required packages

pip install -r requirements.txt

scripts are stored under SDFusion/scripts folder. Run train_ply2shape_multi.sh for multi-gpu training (powered by accelerate package)

GaPartNet dataset is implemented in datasets/gapnet_dataset.py