datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
religious-texts-raw58_books_religious_textThis dataset was made from 58 books on religion. The books were only processed by cleaning scripts and still contain elemets such as page nubers, table of contents, glossaries and other unwanted data.
This was created to establish a tuning workflow with the intention of improving the data later.