[restart] [share link] select language: [swap] [compare scores] [compare models] [map] [release history] [uploads]

OPUS-MT Dashboard

Model Scores (selected model)

IDLanguageBenchmarkOutputbleu
0deu-engflores101-devtestshow 40.2
1deu-engmulti30k_task2_test_2016show 4
2deu-engmulti30k_test_2016_flickrshow 39.6
3deu-engmulti30k_test_2017_flickrshow 40.1
4deu-engmulti30k_test_2017_mscocoshow 31.4
5deu-engmulti30k_test_2018_flickrshow 36.2
6deu-engnewssyscomb2009show 28.5
7deu-engnewstest2009show 26.6
8deu-engnewstest2010show 30.2
9deu-engnewstest2011show 27.5
10deu-engnewstest2012show 29
11deu-engnewstest2013show 32.2
12deu-engnewstest2014show 33.6
13deu-engnewstest2015show 34
14deu-engnewstest2016show 39.9
15deu-engnewstest2017show 35.5
16deu-engnewstest2018show 43.4
17deu-engnewstest2019show 39.2
18deu-engnewstest2020show 39.5
19deu-engnewstestB2020show 37.7
20deu-engtatoeba-test-v2020-07-28show 55.2
21deu-engtatoeba-test-v2021-03-30show 53.7
22deu-engtatoeba-test-v2021-08-07show 52.2
average36.061