| ID | Language | Benchmark | Output | bleu |
|---|---|---|---|---|
| 0 | fra-eng | flores101-devtest | show | 41.7 |
| 1 | fra-eng | multi30k_test_2016_flickr | show | 46.3 |
| 2 | fra-eng | multi30k_test_2017_flickr | show | 48.6 |
| 3 | fra-eng | multi30k_test_2017_mscoco | show | 47.3 |
| 4 | fra-eng | multi30k_test_2018_flickr | show | 41.4 |
| 5 | fra-eng | newsdiscusstest2015 | show | 38.6 |
| 6 | fra-eng | newssyscomb2009 | show | 30.3 |
| 7 | fra-eng | newstest2009 | show | 30.1 |
| 8 | fra-eng | newstest2010 | show | 32.1 |
| 9 | fra-eng | newstest2011 | show | 32.9 |
| 10 | fra-eng | newstest2012 | show | 32.7 |
| 11 | fra-eng | newstest2013 | show | 33.9 |
| 12 | fra-eng | newstest2014 | show | 38 |
| 13 | fra-eng | tatoeba-test-v2020-07-28 | show | 57.9 |
| 14 | fra-eng | tatoeba-test-v2021-03-30 | show | 58.1 |
| 15 | fra-eng | tatoeba-test-v2021-08-07 | show | 58.3 |
| 16 | fra-eng | tico19-test | show | 39.4 |
| average | 41.624 |