I am trying to split a dataset into validation and training set. The dataset is in folders so I have created a function to randomly select some images from each subfolder and that is where I produced a CSV file with all image names in the validation set.
I have tried to use
data = (ImageImageList.from_folder(images)
.split_from_df(df['idx'])
)
And it does not work.
How would you suggest this can be split using the filenames from csv.