Dataset Splits

#1
by stefan-it - opened

Hi @bibwiss , @labusch , @Jrglmn and @cneud ,

I would like to use this awesome dataset for benchmarking encoder models, so I would like to ask if you could accidentally share the exact training, dev and test splits from the paper (or a script that is creating identical splits).

Many thanks in advance and best,

Stefan

Staatsbibliothek zu Berlin - Preußischer Kulturbesitz org

Dear @stefan-it ,

thank you for your request. We have added the data splits for ZEFYS2025 accordingly (https://huggingface.co/datasets/SBB/ZEFYS2025/tree/main/splits). Please let us know if you are also interested in the train/test/dev splits for the other datasets from our paper — we will find a way to make those available as well.

Kind regards,
Sophie

Thanks so much @bibwiss ! The splits are really great to have for ZEFYS2025, so I am fine for now ❤️

stefan-it changed discussion status to closed

Sign up or log in to comment