Towards a fully differentiable digital twin for solar cells : data for machine learning training

Supplementary data for the work conducted and presented in the manuscript: "Towards a fully differentiable digital twin for solar cells". This consists of the data used to train two different machine learning models: one for the prediction of charge carrier generation rates, at the optics module, and one for the prediction of JV related properties, at the electrics module. These are downloadable as two zip files : optics_module.zip for the first model, and electrics_module.zip for the second model. In case of use of the data, please consider citing the article "Towards a fully differentiable digital twin for solar cells".