HiT: Hierarchical Transformers for Unsupervised 3D Shape Abstraction

Aditya Vora, Lily Goli, Andrea Tagliasacchi, Hao Zhang

This repository contains the official implementation of HiT: Hierarchical Transformers for Unsupervised 3D Shape Abstraction.

Installation Steps

1. Clone the Repository:

   git clone --recursive https://github.com/aditya-vora/HiT
   cd HiT

2. Set Up the Conda Environment:

   conda create -n hit python=3.10;
   conda activate hit

3. Install the dependencies inside the conda environment:

   pip install -r requirements.txt

TODO

• Release training code.
• Release data processing scripts.
• Release evaluation scripts.
• Release the pretrained models.
• Testing.

BibTeX

    @article{vora2025hierarchical,
        title={Hierarchical Transformers for Unsupervised 3D Shape Abstraction},
        author={Vora, Aditya and Goli, Lily and Tagliasacchi, Andrea and Zhang, Hao},
        journal={arXiv preprint arXiv:2510.27088},
        year={2025}
    }