This repository contains the training data and pretrained models for the paper "Latent diffusion models for generative precipitation nowcasting with accurate uncertainty quantification". To use the data, clone the repository at https://github.com/MeteoSwiss/ldcast. Unzip the files as follows: Demo files "ldcast-demo-20210622.zip" to the "data" directory Training and evaluation data archive "ldcast-datasets.zip" to the "data" directory Pretrained model archive "models-genforecast.zip" to the "models" directory

Supported by the fellowship ``Seamless Artificially Intelligent Thunderstorm Nowcasts'' from the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT). The hosting institution of this fellowship is MeteoSwiss in Switzerland.