Datasets:
Dataset Viewer
View in Dataset Viewer
Viewer
The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
Entity Disambiguation datasets as provided in the GENRE repo. The dataset can be used to train and evaluate entity disambiguators.
The datasets can be imported easily as follows:
from datasets import load_dataset
ds = load_dataset("boragokbakan/entity_disambiguation", "aida")
Available dataset names are:
blinkace2004aidaaquaintblinkcluewebmsnbcwiki
Note: As the BLINK training set is very large in size (~10GB), it is advised to set streaming=True when calling load_dataset.
- Downloads last month
- 2