2 ms·
They also test a Lucene based model, which is much worse on Top1, but similar to SWIM for Top 3 and Top 10. Also, BLEU is a fairly widely used metric. I assume
by nl 10y ago
They also test a Lucene based model, which is much worse on Top1, but similar to SWIM for Top 3 and Top 10.
Also, BLEU is a fairly widely used metric. I assumed they referenced the paper though?