VQNet – Quantum Machine Learning SDK
Supports Both Quantum and Classical Model Building and Training,Developed by Origin Quantum, VQNet enables efficient simulation and training of neural networks.
Get Started with VQNet ->
VQNet supports the construction and training of both quantum machine learning and classical machine learning models. VQNet integrates built-in quantum computing modules, enabling efficient simulation training of quantum network models.
It also incorporates the QPanda module, supporting model training and inference on real quantum computers. The documentation section provides abundant examples for learning purposes, including variational quantum classifiers, hybrid quantum-classical neural network models and U-Net models, unsupervised learning—quantum clustering, and more.
VQNet supports users to conduct research and development of quantum machine learning in a variety of hardware and operating system environments. Whether using CPU or GPU for quantum computing simulation or calling real quantum chips through origin Quantum Cloud Service, VQNet can provide seamless support. Currently, VQNet is compatible with python3.9, python3.10, and python3.11 versions of Windows, Linux, and macOS systems.
VQNet uses Python as the front-end language, provides a function interface similar to PyTorch, and can freely choose a variety of computing backends to implement the automatic differentiation function of classical quantum machine learning models. The framework has built-in: 100+ commonly used Tensor computing interfaces, 100+ quantum variational circuit computing interfaces, and 50+ classical neural network interfaces. These interfaces cover the complete development process from classical machine learning to quantum machine learning, and will be continuously updated.
For users who need real quantum chip experiments, VQNet integrates the original pyQPanda interface, and combines the efficient scheduling capabilities of the original Sinan to achieve fast quantum circuit simulation calculations and real chip operation.For local computing needs, VQNet provides a quantum machine learning programming interface based on CPU or GPU, and uses automatic differentiation technology to perform quantum variational circuit gradient calculations, which is significantly faster than traditional parameter drift methods (such as Qiskit).
VQNet is not only a powerful development tool, but also widely used in multiple projects within the company, including power optimization, medical data analysis, image processing and other fields. In order to help users get started quickly, VQNet provides a variety of scenarios ranging from basic tutorials to advanced applications in the official website and API online documentation. These resources enable users to easily understand how to use VQNet to solve practical problems and quickly build their own quantum machine learning applications.
Uses VQNet to build a quantum circuit for binary classification by encoding inputs and optimizing parameters to distinguish odd and even.
Builds a simple neural network with VQNet to show its ease of use and explore quantum applications.
This example integrates pyqpanda with VQNet to partially quantize the classical Unet model, forming a Quantum Unet (QUnet) suitable for quantum data.
Quantum clustering is a typical unsupervised method that groups samples based on similarity, helping uncover hidden patterns in data.