How should data be prepared and placed when directly using the model_SARcarl.hdf5 model?

[lzux]

When using the model_SARcarl.hdf5 model directly for prediction, does s2_cloudy have to include all 13 bands? Do the sizes of s1 and s2_cloudy also have to be exactly the same? In addition, what format should be set for the names of subfolders and images in the input_data_folder path? Thanks~

[ameraner]

Hi @lzux , yes to the first two questions. The folders should be called s1 s2_cloudFree s2_cloudy, basically as defined in your datasetfilelist.csv.