VIPR Reflectometry Model: Neural Spline Flow for Neutron Reflectometry Inference (MARIA Dataset)

This dataset provides a pre-trained Neural Spline Flow (NSF) model for probabilistic inference in neutron reflectometry, developed within the VIPR framework.

The model estimates posterior distributions of thin-film parameters (e.g., thickness, SLD, roughness) directly from reflectivity curves, enabling fast uncertainty quantification and detection of multimodal solutions. It is configured for data from the MARIA reflectometer and corresponds to the probabilistic workflow presented in the associated VIPR publication .

Contents: model weights and configuration file.<br>
Usage: designed for execution within VIPR via YAML-defined inference workflows.