QSPLIT: A GUI-based testing framework for quantum split learning

The implementation of Quantum Neural Networks (QNN) is challenging because it requires specialized knowledge of quantum mechanics, posing a significant obstacle for developers and limiting broader adoption. In this paper, we present QSPLIT, a GUI-based testing framework for efficient testing of QNN architectures within quantum split learning environments. QSPLIT allows developers to provide a subset of QNN components as target code for validation, while automatically generating the remaining components as dummy code to construct a complete architecture. The framework executes parallel split learning across multiple target–dummy code combinations, providing GUI-based tools for real-time log tracking, result comparison, and code export. By integrating these functionalities, QSPLIT streamlines the development and validation of QNN models. Its effectiveness was demonstrated through experiments on MedNIST, a medical imaging dataset.