Is there unsplit version?

#2
by murodbek - opened

Hello guys. Thank you for great dataset for Kazakh language.

We have been experimenting with it and found out that some entries is split into sentences.
image.png
This is the distribution of number of sentences across different splits. Tokenized to sentences using qaznltk

Is there any way we can restore the whole entries? Thanks in advance.

murodbek changed discussion title from Is there not clipped version? to Is there unsplit version?

Sign up or log in to comment