Labels.csv in planet_sample has index column

labels.csv in planet_sample datasets (PLANET = f'{URL}planet_sample') on has the index 0, 1,... as the first column. This does not seem consistent with the convention. (For example the defaults are fn_col=0 and label_col=1 in ImageDataBunch.from_csv. Perhaps index=False was missing in to_csv when creating this?

1 Like

Fixed now.