PartSAM: A Scalable Promptable Part Segmentation Model Trained on Native 3D Data

Zhe Zhu¹, Le Wan², Rui Xu³, Yiheng Zhang⁴, Honghua Chen⁵, Zhiyang Dou³, Cheng Lin⁶, Yuan Liu^2†, Mingqiang Wei^1†
† Corresponding authors

¹ Nanjing University of Aeronautics and Astronautics ² Hong Kong University of Science and Technology ³ The University of Hong Kong ⁴ National University of Singapore ⁵ Lingnan University ⁶ Macau University of Science and Technology

Installation

1. Install the required environment

conda create -n PartSAM python=3.11 -y
conda activate PartSAM
# PyTorch 2.4.1 with CUDA 12.4
pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu124
pip install lightning==2.2 h5py yacs trimesh scikit-image loguru boto3
pip install mesh2sdf tetgen pymeshlab plyfile einops libigl polyscope potpourri3d simple_parsing arrgh open3d safetensors
pip install hydra-core omegaconf accelerate timm igraph ninja
pip install torch-scatter -f https://data.pyg.org/whl/torch-2.4.1+cu124.html
apt install libx11-6 libgl1 libxrender1
pip install vtk

2. Install other third-party modules (torkit3d and apex) following Point-SAM

3. Install the pretrained model weight

pip install -U "huggingface_hub[cli]"
huggingface-cli login
huggingface-cli download Czvvd/PartSAM --local-dir ./pretrained

Usage

# Modify the config file to evaluate your own meshes
python evaluation/eval_everypart.py

TODO

• Release inference code of PartSAM
• Release the pre-trained models
• Release training code and data processing script

Acknowledgement

Our code is based on these wonderful works:

• Point-SAM
• PartField
• SAMPart3D
• SAMesh
• Find3D

We thank the authors for their great work！

Citation

@article{zhu2025partsam,
  title={PartSAM: A Scalable Promptable Part Segmentation Model Trained on Native 3D Data}, 
  author={Zhe Zhu and Le Wan and Rui Xu and Yiheng Zhang and Honghua Chen and Zhiyang Dou and Cheng Lin and Yuan Liu and Mingqiang Wei},
  journal={arXiv preprint arXiv:2509.21965},
  year={2025}
}