Could you please provide more details about the dataset format, including the folder structure, file types, and naming conventions?

[ljphr]

Could you please provide more details about the dataset format, including the folder structure, file types, and naming conventions?
Best regards

[Mnostalgia]

I modified the dataset part of the readme a bit for you.
My complete dataset consisted of rgb, ndvi, and labels folders, with images with JPG, TIF, and png extensions under them.
However, if necessary, you can modify the folder name and extension name in line 60 of util/MF_dataset.py to suit your dataset.

For the same pair of datasets, rgb, ndvi, and label all use the same name.
And in train.txt and test.txt, just write the name of the dataset you want to train and test according to the desired ratio without the extension name.