metadata

annotations_creators:
  - expert-generated
language_creators:
  - found
languages:
  - en
licenses:
  - mit
multilinguality:
  - monolingual
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - structure-prediction
task_ids:
  - structure-prediction-other-acronym-identification

XL-WiC

Huggingface dataset for the XL-WiC paper https://www.aclweb.org/anthology/2020.emnlp-main.584.pdf. Please refer to the official website for more information.

Languages

XL-WiC provides training data in:

en (English)
fr (French)
de (German)
it (Italian)

and dev and test data in:

fr (French)
de (German)
it (Italian)
bg (Bulgarian)
zh (Chinese)
hr (Croatian)
da (Danish)
nl (Dutch)
et (Estonian)
fa (Farsi)
ja (Japanesse)
ko (Korean)

Configurations

When loading one of the XL-WSD datasets one has to specify the training language and the target language (on which dev and test will be performed). For example, we can load the dataset having English as training language and Italian as target language as follows:

from datasets import load_dataset
dataset = load_dataset('pasinit/xlwic', 'en_it')