OPUS-MT Dashboard

blue = OPUS-MT / Tatoeba-MT models, grey = external models, purple = user-contributed
render chart with [gd] [plotly]

ID	Benchmark (bleu)	Output	OPUS-MT	bleu	external	bleu	Diff
0	flores101-devtest	compare	ara-eng/opus...2022-03-09	42.6	facebook/nllb-200-3.3B	41.1	1.5
1	flores200-devtest	compare	ara-eng/opus...2022-03-09	42.8	facebook/nllb-200-3.3B	41.3	1.5
2	tatoeba-test-v2020-07-28	compare	ara-eng/opus...2022-03-09	47.5	facebook/nllb-200-3.3B	49.8	-2.3
3	tatoeba-test-v2021-03-30	compare	ara-eng/opus...2022-03-09	47.3	facebook/nllb-200-3.3B	49.7	-2.4
4	tatoeba-test-v2021-08-07	compare	ara-eng/opus...2022-03-09	47.3	facebook/nllb-200-3.3B	49.7	-2.4
5	tico19-test	compare	ara-eng/opus...2022-03-09	44.4	facebook/nllb-200-3.3B	43.5	0.9
		average		45.3		45.8	-0.5