length train = 1259113
length test = 3117
the train and test sets are balanced. Where there is only one sample in the full_set, the sample is assigned to the train_set
# templates train_set = 191
# templates test_set = 189
Text-To-SPARQL dataset for the lexicographical data ontology module on Wikidata.
The dataset was generated with 78 Templates that were populated with Data of the Wikidata Knowledge Graph. The dataset contains approximately 1.25M rows.
The dataset has 3 columns:
- utterance: A natural language utterance
- template_name: The name of the template that generated the tuple
- template: The populated template
