| ID | Benchmark (bleu) | Output | OPUS-MT | bleu | external | bleu | Diff |
|---|---|---|---|---|---|---|---|
| 0 | flores200-devtest | compare | mul-mul/opus...2024-08-17 | 1.5 | facebook/nll...illed-1.3B | 10.9 | -9.4 |
| 1 | tatoeba-test-v2020-07-28 | compare | mul-mul/opus...2024-08-17 | 6.8 | facebook/nll...illed-600M | 25 | -18.2 |
| 2 | tatoeba-test-v2021-03-30 | compare | mul-mul/opus...2024-08-17 | 7 | facebook/nll...illed-600M | 24.9 | -17.9 |
| 3 | tatoeba-test-v2021-08-07 | compare | mul-mul/opus...2024-08-17 | 6.8 | facebook/nll...illed-600M | 25.2 | -18.4 |
| average | 5.5 | 21.5 | -16.0 |