Data for "Predicting high-intensity femtosecond x-ray exposure on protein crystals with simulated x-ray emission spectra"

Data for "Interpretable neural network predictions of high-intensity femtosecond x-ray free-electron laser pulses using sulfur K-shell emission spectra" The data here is organized as follows. With dataset-1 we train a model to predict x-ray pulse fluence values. With dataset-4 we train a model to predict both fluence and pulse duration. The main directory has scripts for processing and augmenting the raw data in data_processing.npy. The models are trained using the scripts ds1_model_training.npy and ds4_model_training.npy. There are additional scripts and results for computing the interpretation scores based on gradients and shap ds1_interpretation_score.ipynb and ds4_interpretation_score.ipynb. The scripts for training a model on occluded data are in ds1_model_training_occlusion.ipynb. The decoupled training of fixed fluence and fixed pulse duration are in ds4_model_training_decouple.ipynb. Finally, ds4_hypertunner.ipynb shows how we ran the hyper tunner to select a good model architecture.