CoolFace
Datasetpublic

google/wiki40b

Dataset Card for "wiki40b" Dataset Summary Clean-up text for 40+ Wikipedia languages editions of pages correspond to entities. The datasets have train/dev/test splits per language. The dataset is cleaned up by page filtering to remove disambiguation pages, redirect pages, deleted pages, and non-entity pages. Each example contains the wikidata id of the entity, and the full Wikipedia article after page processing that removes non-content sections and structured… See the full description on the dataset page: https://huggingface.co/datasets/google/wiki40b.

sourceHugging Faceupdated 3y agoView on Hugging Face
37likes11kdownloads
../
filetest-00000-of-00001.parquet5.7 MBdownload
filetrain-00000-of-00001.parquet105.0 MBdownload
filevalidation-00000-of-00001.parquet5.9 MBdownload

google/wiki40b · main · files are served by the source, never re-hosted here