I am trying to do multi-class classification using AID dataset , for remote sensing. It has 30 classes with 350+ images each. Finding it hard to figure out how to create databunch to access the images. Do i have to create seperate csv labels? Folders are already labeled individually.
You probably want to do something like
ImageList.from_folder(’/content/AID’)
.split_by_files(list_of_files_for_validation)
.label_from_folder()