This repository contains the code for the interpretable ML-based radiation emulator for ICON. The corresponding paper is currently under review in the Journal of Geophysical Research - Machine Learning and Computation
If you want to use this repository, start by executing
conda env create -f environment.yml
conda activate hafner_ml_rad
- evaluation contains some functions for prediction and evaluation
- models contains the NN architecture including preprocessing layer
- nn_config contains the configuration of all NNs
- plotter contains plotting functions
- preprocessing contains the normalization file and data loader
- utils contains some helper functions
- train_coarse_levante.py contains the training script
- eval_coarse_levante.py contains the evaluation script
- eval_coarse_levante.ipynb contains some further evaluation