Pretrained models and results for "Thunderstorm nowcasting with deep learning: a multi-hazard data fusion model"

<p>This dataset contains the pretrained model weights and precomputed results for the paper &quot;Thunderstorm nowcasting with deep learning: a multi-hazard data fusion model&quot; submitted to <em>Geophysical Research Letters</em>. A preprint of the paper can be found at <a href="https://arxiv.org/abs/2211.01001">https://arxiv.org/abs/2211.01001</a>.</p> <p>The ML code can be found at <a href="https://github.com/MeteoSwiss/c4dl-multi">https://github.com/MeteoSwiss/c4dl-multi</a>. Download all the files here and extract the contents to the following subdirectories in the ML code directory:</p> <ul> <li>Results (<a href="https://zenodo.org/api/files/d4829f50-55fd-4d86-b875-7f2b91dba74f/c4dl-results-lightningdl.zip?versionId=54046830-4c7e-48c6-af42-d6d5606af86b">c4dl-results-lightningdl.zip</a>) -&gt; results/</li> <li>Pretrained models (<a href="https://zenodo.org/api/files/d4829f50-55fd-4d86-b875-7f2b91dba74f/c4dl-models-lightningdl.zip?versionId=364bca7c-e6ad-4ed9-9264-57c759ea0ac6">c4dl-models-lightningdl.zip</a>) -&gt; models/</li> <li>If you want to train models, data (<a href="https://zenodo.org/api/files/0cfc0cf4-755a-4618-8341-39b107a3901c/c4dl-patches-2020-additional.zip">c4dl-patches-2020-additional.zip</a>) -&gt; data/2020/</li> </ul> <p>Additionally, you will need the datasets from <a href="https://zenodo.org/deposit/6802292">this Zenodo archive</a>. Follow the instructions there for downloading.</p>