FLORES
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Flores
40.91BLEU Score
97
FLORES
89.07Average Score
47
Flores-200 Romance group xx->en (test)
42.88BLEU
46
FLORES
88.9Score
43
FLORES (test)
64.4chrF2 Score
42
FLORES-200 (devtest)
5.67Delta CHRF++
39
FLORES xx→en (test)
-65.05Score (de→en)
38
FLORES-200 XX ⇔ XX
88.74XCOMET-XXL Score
30
FLORES-200 ZH ⇔ XX
90.3XCOMET-XXL
30
FLORES Medium Resource
38.33BLEU (En→X)
27
FLORES High Resource
38.34En->X BLEU
27
FLORES+
24Score (es)
27
FLORES EN-ZH
90.08COMETkiwi
26
FLORES+ (test)
69.2Vocabulary Utilisation
24
FLORES Low Resource
39.32BLEU (En->X)
24
FLORES-200
79.23COMET
23
FLORES-200
91.42GRF
21
FLORES-200 XX ⇔ XX 2022
87.73XCOMET-XXL
17
FLORES-200 EN ⇔ XX 2022
94.13XCOMET-XXL
17
FLORES-200 ZH ⇔ XX 2022
0.898XCOMET-XXL
17
Flores-101 (val test)
46.8CHRF
17
FLoRes+ En-YY (total)
44.6ChrF++
16
FLoRes+ En-YY mid resource level
46.5ChrF++
16
FLoRes+ En-YY, high resource level
58.6ChrF++
16
FLoRes+ XX-En, low resource level
54.6ChrF++
16