OPUS-MT Dashboard

Language pair: [eng-fra] all languages
Models: [all models] [OPUS-MT] [external] [compare]
Selected: Tatoeba-MT-models/fin-eng/opusTCv20210807+nopar+ft95-sepvoc_transformer-align_2022-01-28
Benchmark: all benchmarks [download]
Evaluation metric: bleu [spbleu][chrf][chrf++][comet]
Chart Type: [barchart] [heatmap]

orange = OPUS-MT, blue = Tatoeba-MT models, red = HPLT-MT models
green = student models, grey = external models, purple = user-contributed
render chart with [gd] [plotly]

Model Scores (selected model)

ID	Language	Benchmark	Output	bleu
0	fin-eng	flores101-devtest	show	33.3
1	fin-eng	flores200-devtest	show	33.3
2	fin-eng	newstest2015	show	28.5
3	fin-eng	newstest2016	show	32
4	fin-eng	newstest2017	show	34.9
5	fin-eng	newstest2018	show	25.3
6	fin-eng	newstest2019	show	30.8
7	fin-eng	newstestB2016	show	26.1
8	fin-eng	newstestB2017	show	29.4
9	fin-eng	ntrex128	show	28.6
10	fin-eng	tatoeba-test-v2020-07-28	show	54.9
11	fin-eng	tatoeba-test-v2021-03-30	show	54.9
12	fin-eng	tatoeba-test-v2021-08-07	show	54.8
			average	35.908