Alleviating the Inequality of Attention Heads for Neural Machine Translation

4758

Knowledge Triples