Efficient long-distance relation extraction with DG-SpanBERT

Benchmark Model Rank Results
relation-extraction-on-tacredDG-SpanBERT-largeF1: 71.5