Machine Translation on Law (test)
61.55BLEUSK-MT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SK-MTmode=SK-MT2, m=16, k=22024.06 | 61.55 | 87.68 | |
| Vanilla kNN-MT2024.06 | 61.16 | 87.46 | |
| SK-MTmode=SK-MT1, m=2, k=12024.06 | 60.43 | 87.46 | |
| kNN-MT-DR2024.06 | 60.1 | 87.39 | |
| kNN-MT-ARalpha=0.252024.06 | 59.89 | 87.57 | |
| Faster kNN-MT2024.06 | 58.97 | 87.36 | |
| kNN-MT-ARalpha=0.52024.06 | 58.83 | 87.5 | |
| kNN-MT-ARalpha=0.752024.06 | 51.11 | 86.65 | |
| Base NMT2024.06 | 45.48 | 85.66 |