This repository contains the DNN model construction, training loop code and inference code. The code for the data generation is in generate_dataset.ipynb. It contains function that allows you to check the shape of cross section and the parameters used.
Use the jupyter notebook experiment.ipynb to perform DNN model inferences on separable potential and nucleon-nucleon data in NN-Online. The pretrained models are saved in generalization folder. If you want to skip the generation of validation dataset, you may upload the separable validation dataset in inference_data. It is already in the experiment.ipynb code.
If you wish to use this code in your work, please cite our papers: PhysRevD.102.01602 and Few-Body Syst 62, 52 (2021).