Installation Guide#

This chapter explains how to install Kaiwu-PyTorch-Plugin and its dependencies.

1. System Requirements#

Before installation, make sure your system meets the following requirements:

Dependency

Version Required

Python

3.10

PyTorch

2.7.0

NumPy

2.2.6

Kaiwu SDK

v1.2.0+

Check your Python version:

 python --version
 # or
 python3 --version

If you need to install Python 3.10, visit the Python 3.10 download page <https://www.python.org/downloads/release/python-31011/ >_.

2. Installing Kaiwu-PyTorch-Plugin#

2.1 Create and Activate Environment#

We recommend creating an isolated Python environment with conda:

 # Create a new environment
 conda create -n quantum_env python=3.10

 # Activate the environment
 conda activate quantum_env

2.2 Clone the Repository#

Clone the project from GitHub to your local machine:

 git clone https://github.com/QBoson/Kaiwu-pytorch-plugin.git
 cd kaiwu-pytorch-plugin

2.3 Install Dependencies#

Install project dependencies:

 pip install -r requirements/requirements.txt

2.4 Install the Plugin#

 pip install .

3. Install Kaiwu SDK (Required)#

Kaiwu-PyTorch-Plugin relies on Kaiwu SDK for quantum-computing capabilities; you must install Kaiwu SDK separately.

3.1 Obtain the SDK#

  1. Visit the Kaiwu SDK download page <https://platform.qboson.com/sdkDownload >_

  2. Download the SDK package for your system

  3. Follow the Kaiwu SDK installation guide to complete installation

3.2 Configure Credentials#

After installation, configure your SDK credentials:


 User ID: <your-user-id>
 SDK Token: <your-sdk-token>

Note

Replace the placeholders with your actual credentials, which can be obtained from the Kaiwu SDK page on the QBoson platform.

4. Verify Installation#

After installation, run the following code to verify success:

 # 验证 PyTorch
 import torch
 print(f"PyTorch version: {torch.__version__}")



 # 验证 Kaiwu SDK
 import kaiwu
 import numpy as np
 from kaiwu.classical import SimulatedAnnealingOptimizer
 opt = SimulatedAnnealingOptimizer()
 mat = np.array([[1, -1], [-1, 1]])
 result = opt.solve(mat)
 print(f"Kaiwu SDK version: {kaiwu.__version__}")
 print(result)

 # 验证 Kaiwu-PyTorch-Plugin
 from kaiwu.torch_plugin import RestrictedBoltzmannMachine
 print("Kaiwu-PyTorch-Plugin imported successfully!")
 # 简单测试
 rbm = RestrictedBoltzmannMachine(num_visible=10, num_hidden=5)
 print(f"RBM created with {rbm.num_visible} visible and {rbm.num_hidden} hidden units")

If no errors appear, installation is complete. You can now build models and validate them with classical simulators; once satisfied, switch to a quantum backend to leverage real quantum resources.

5. Obtain Quantum-Computer Access#

To experience genuine quantum computing power, you need access to a quantum computer:

  1. Register an account on the QBoson platform

  2. Contact official staff through the platform to obtain real-machine quota

Note

While awaiting real-machine access, you can develop and test with simulators. Kaiwu SDK provides several classical optimizers (e.g. simulated-annealing optimizer) as classical substitutes for quantum samplers.

6. Development Environment Setup (Optional)#

If you plan to contribute to the plugin, install the development dependencies:

 pip install -r requirements/devel.txt

Run tests:

 # Run all tests
 pytest tests/

 # Run specific tests
 pytest tests/test_rbm.py

Code-style check:

 pylint kaiwu/

7. FAQ#

Q: Installation reports incompatible Python version?#

A: Kaiwu-PyTorch-Plugin currently requires Python 3.10. Create a Python 3.10 environment with conda:

 conda create -n quantum_env python=3.10
 conda activate quantum_env

Q: Cannot import kaiwu.torch_plugin?#

A: Please ensure that:

  1. the environment is activated (conda activate quantum_env)

  2. Kaiwu SDK is correctly installed

  3. Kaiwu-PyTorch-Plugin is correctly installed (pip install .); check with

Q: How do I update to the latest version?#

A: Enter the project directory and pull the latest code:

 cd kaiwu-pytorch-plugin
 git pull origin main
 pip install .
On this page