P.808/docs/prep_p835.md

[heshulin]

P.808/docs/prep_p835.md
In this file, what are the speech in the three csvs? Are they clean speech, predicted speech, and training set? Some csvs need to have a score. Is this score manually annotated by myself?