| ID | Language | Benchmark | Output | bleu |
|---|---|---|---|---|
| 0 | deu-ces | flores101-devtest | show | 23.3 |
| 1 | deu-ces | multi30k_test_2016_flickr | show | 26.3 |
| 2 | deu-ces | multi30k_test_2018_flickr | show | 24 |
| 3 | deu-ces | newssyscomb2009 | show | 21.6 |
| 4 | deu-ces | newstest2009 | show | 20.6 |
| 5 | deu-ces | newstest2010 | show | 22.2 |
| 6 | deu-ces | newstest2011 | show | 20.6 |
| 7 | deu-ces | newstest2012 | show | 20.3 |
| 8 | deu-ces | newstest2013 | show | 22.4 |
| 9 | deu-ces | newstest2019 | show | 20.7 |
| 10 | deu-ces | tatoeba-test-v2021-08-07 | show | 41.7 |
| average | 23.973 |