FastSA

About

Fast Simulation Annealing for QUBO Problem with GPU and Batch Support by C++ and libtorch.

Citation

The paper QAMA implement the code. If this helps, please cite:

@misc{du2025qamascalablequantumannealing,
      title={QAMA: Scalable Quantum Annealing Multi-Head Attention Operator for Deep Learning}, 
      author={Peng Du and Jinjing Shi and Wenxuan Wang and Yin Ma and Kai Wen and Xuelong Li},
      year={2025},
      eprint={2504.11083},
      archivePrefix={arXiv},
      primaryClass={quant-ph},
      url={https://arxiv.org/abs/2504.11083}, 
}

Installation

1. Switch to your environment:

conda activate your_environment

Or you can create a temporary environment through our development environment:

conda create -n FastSA python==3.13.5
conda activate FastSA
pip install -r requirements-dev.txt

Examples

Runtime Experiments

Note: Limitations and Future Improvements

Acknowledgements

PennyLane Lightning makes use of the following libraries and tools, which are under their own respective licenses: