Machine Translation on newstest En-Ru 2015 (test)
24.3BLEUState-of-the-art Non-Neural Approach
Evaluation results
| Method | Links | |
|---|---|---|
| State-of-the-art Non-Neural ApproachApproach=Non-Neural2016.03 | 24.3 | |
| Bi-scale Character-level NMT (Row m)Target=Char, Depth=2, Attention=Concatenated, Model=Bi-S, Ensemble=true2016.03 | 23.75 | |
| Base Character-level NMT (Row l)Target=Char, Depth=2, Attention=Concatenated, Model=Base, Ensemble=true2016.03 | 23.51 | |
| Base BPE-level NMT (Row k)Target=BPE, Depth=2, Attention=Concatenated, Model=Base, Ensemble=true2016.03 | 22.96 | |
| Base Character-level NMT (Row l)Target=Char, Depth=2, Attention=Concatenated, Model=Base, Ensemble=false2016.03 | 21.1 | |
| Bi-scale Character-level NMT (Row m)Target=Char, Depth=2, Attention=Concatenated, Model=Bi-S, Ensemble=false2016.03 | 20.73 | |
| Base BPE-level NMT (Row k)Target=BPE, Depth=2, Attention=Concatenated, Model=Base, Ensemble=false2016.03 | 19.72 |