stop iterating through dataset after max_samples is reached and read language filter columns first (with license in the beginning now)

#7
by KT313 - opened

saves a loooot of time generating the train split, e.g. if you only need 100k samples from the python subset or so

Ready to merge
This branch is ready to get merged automatically.

Sign up or log in to comment