Stack: Python, PyTorch, automatic differentiation, quantum computing.
Qadence is a scientific software package, written mainly in Python, for building and executing quantum computing programs with a simple yet powerful interface. Qadence focuses on easiness to use, flexibility and performance by integrating with other open-source quantum computing simulators (such as PyQTorch to which I also contributed).
Qadence mainly targets applications in the (quantum) machine learning domain. Therefore, automatic
differentiation plays an integral role in its design. In particular, it extends
the popular PyTorch deep learning framework making quantum program written in Qadence automatically
differentiable by default by simply using PyTorch autograd differentiation engine.
Qadence was the resulting effort of over one year of work of my team at PASQAL. I was the main driver behind this project, steering the project, contributing to the most important design choices and implementations, and maintaining the roadmap for future of the package.
Qadence is now an open-source project which can be found on GitHub. We recently published a review paper with many details on its implementation and features.