Each text file is a list of sequence IDs. We sampled from the original train set samples, that has count annotations, which is 1780 samples total.
count