FractalCloud

FractalCloud Artifact Repository

This repository provides the Docker image and pretrained models for the HPCA’26 paper:

FractalCloud: A Fractal-Inspired Architecture for Efficient Large-Scale Point Cloud Processing

The official implementation (including full source code, training pipelines, and evaluation scripts) is available at:
👉 https://github.com/Yuzhe-Fu/FractalCloud

Contents

• Docker image for reproducing all experiments in the paper
• Pretrained models for classification and segmentation tasks
• Fully packaged environment with all dependencies included

Usage

Please refer to the official repository for instructions on:

• environment setup
• dataset preparation
• inference, training, and finetuning
• experiment reproduction

All steps follow the procedure described in the paper and the official codebase.

Citation

If you find this repository useful in your research, please cite:

@inproceedings{fu2026fractalcloud,
  title={FractalCloud: A Fractal-Inspired Architecture for Efficient Large-Scale Point Cloud Processing},
  author={Fu, Yuzhe and Zhou, Changchun and Ye, Hancheng and Duan, Bowen and Huang, Qiyu and Wei, Chiyue and Guo, Cong and Li, Hai Helen and Chen, Yiran},
  booktitle={2026 IEEE International Symposium on High Performance Computer Architecture (HPCA)},
  pages={1--15},
  year={2026},
  organization={IEEE}
}