Open gordom6 opened 4 years ago
Data is here: https://github.com/adapt-sjtu/commonsense-locatednear/blob/master/data The "1"-labeled pairs in the 500 test pairs are fairly good. The "1"-labeled pairs in the 5000 pairs are inconsistent.
Not worth the time currently to extract ~200-300 relations, may use this as a student intro later.
https://github.com/adapt-sjtu/commonsense-locatednear