Machine Translation on Tatoeba (test)
30.2SacreBLEU Score (Da)TEnc/lang
Evaluation Results
| Method | Links | ||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TEnc/langModel=One-model-per-language2025.09 | 30.2 | 0.73 | 23.5 | 0.71 | 28.2 | 0.69 | 14 | 0.6 | 24.3 | 0.72 | 36.2 | 0.8 | 1.6 | 0.46 | 12.7 | 0.66 | 30.8 | 0.77 | 28.8 | 0.73 | 19.2 | 0.68 | 22.7 | 0.69 | |
| TEnc/langModel=TEnc/lang2025.09 | 30.2 | 0.73 | 23.5 | 0.71 | 28.2 | 0.69 | 14 | 0.6 | 24.3 | 0.72 | 36.2 | 0.8 | 1.6 | 0.46 | 12.7 | 0.66 | 30.8 | 0.77 | 28.8 | 0.73 | 19.2 | 0.68 | 22.7 | 0.69 | |
| TETModel=Ours2025.09 | 29.5 | 0.72 | 20.1 | 0.67 | 22.8 | 0.65 | 13.1 | 0.58 | 36.1 | 0.76 | 29.2 | 0.74 | 10.2 | 0.59 | 9.8 | 0.65 | 23.1 | 0.71 | 24.2 | 0.69 | 22.7 | 0.69 | 21.9 | 0.68 | |
| TETModel=TET (Ours)2025.09 | 29.5 | 0.72 | 20.1 | 0.67 | 22.8 | 0.65 | 13.1 | 0.58 | 36.1 | 0.76 | 29.2 | 0.74 | 10.2 | 0.59 | 9.8 | 0.65 | 23.1 | 0.71 | 24.2 | 0.69 | 22.7 | 0.69 | 21.9 | 0.68 | |
| TEnc/allModel=One-model-for-all2025.09 | 27.7 | 0.72 | 17.1 | 0.66 | 16.9 | 0.62 | 10 | 0.55 | 26.6 | 0.73 | 23.7 | 0.72 | 11.2 | 0.62 | 9.8 | 0.63 | 16.3 | 0.66 | 20.9 | 0.68 | 22 | 0.69 | 18.4 | 0.66 | |
| TEnc/allModel=TEnc/all2025.09 | 27.7 | 0.72 | 17.1 | 0.66 | 16.9 | 0.62 | 10 | 0.55 | 26.6 | 0.73 | 23.7 | 0.72 | 11.2 | 0.62 | 9.8 | 0.63 | 16.3 | 0.66 | 20.9 | 0.68 | 22 | 0.69 | 18.4 | 0.66 | |
| TET-RndModel=Random initialization variant2025.09 | 26.1 | 0.71 | 18.8 | 0.66 | 19.7 | 0.63 | 13.2 | 0.57 | 32.9 | 0.75 | 26.7 | 0.72 | 8.6 | 0.58 | 10.3 | 0.64 | 23.1 | 0.7 | 22.6 | 0.68 | 23.4 | 0.7 | 20.5 | 0.67 | |
| TET-RndModel=TET-Rnd2025.09 | 26.1 | 0.71 | 18.8 | 0.66 | 19.7 | 0.63 | 13.2 | 0.57 | 32.9 | 0.75 | 26.7 | 0.72 | 8.6 | 0.58 | 10.3 | 0.64 | 23.1 | 0.7 | 22.6 | 0.68 | 23.4 | 0.7 | 20.5 | 0.67 |