PyQuafu

Introduction

PyQuafu is designed for users to construct, compile, and execute quantum circuits on quantum devices on Quafu using Python. With PyQuafu, you can interact with various real quantum backends provided by the experimental group from Quafu.

Installation

Install via PyPI

You can install PyQuafu directly from PyPI:

pip install pyquafu

Build from Source

Alternatively, you can build PyQuafu from the source:

pip install .

Development Installation

For developers, it is recommended to use the editable installation below, which automatically includes development dependencies (pytest, pre-commit):

pip install -e ".[dev]"

To run unit tests, you can use:

python -m pytest tests

Graphviz Dependency

If you need to visualize Directed Acyclic Graphs (DAGs), ensure that the Graphviz software is installed on your system. Refer to the graphviz · PyPI page for installation guidance.

GPU Support

To install PyQuafu with GPU-based circuit simulation, you need to build from the source and ensure that the CUDA Toolkit is installed. Use the following command to install the GPU version:

python setup.py install -DUSE_GPU=ON

If you also have cuQuantum installed, you can install PyQuafu with cuQuantum support:

python setup.py install -DUSE_GPU=ON -DUSE_CUQUANTUM=ON

Documentation

For detailed documentation about usage, please visit the PyQuafu documentation website.

Note for Apple Silicon Mac Users

If you encounter the error "illegal hardware instruction" on an Apple silicon Mac, ensure that you have updated to the arm64 version of Anaconda. See this issue for more details.

Examples

Quantum Reinforcement Learning

This example demonstrates how quantum reinforcement learning interacts with Quafu to solve the CartPole environment. For more details, refer to the quantum-RL-with-quafu repository.

Author

This project is developed by the quantum cloud computing team at the Beijing Academy of Quantum Information Sciences.