OPUS-MT Dashboard

Language pair: [eng-fra] all languages
Models: [all models] [OPUS-MT] [external] [compare]
Selected: Tatoeba-MT-models/sla-heb/opusTCv20210807_transformer-big_2022-09-15
Benchmark: all benchmarks [download]
Evaluation metric: bleu [spbleu][chrf][chrf++][comet]
Chart Type: [barchart] [heatmap]

ID	Language	Benchmark	Output	bleu
0	bel-heb	flores101-devtest	show	9.6
1	bel-heb	flores200-devtest	show	9.2
2	bel-heb	ntrex128	show	13.5
3	bul-heb	flores101-devtest	show	20.9
4	bul-heb	flores200-devtest	show	20.3
5	bul-heb	ntrex128	show	16
6	ces-heb	flores101-devtest	show	20.2
7	ces-heb	flores200-devtest	show	19.6
8	ces-heb	ntrex128	show	15.9
9	pol-heb	flores101-devtest	show	16
10	pol-heb	flores200-devtest	show	15.5
11	pol-heb	ntrex128	show	13.8
12	pol-heb	tatoeba-test-v2021-03-30	show	38.7
13	pol-heb	tatoeba-test-v2021-08-07	show	38.7
14	rus-heb	flores101-devtest	show	19.5
15	rus-heb	flores200-devtest	show	18.9
16	rus-heb	ntrex128	show	13.3
17	rus-heb	tatoeba-test-v2021-08-07	show	38.4
18	slv-heb	flores101-devtest	show	18
19	slv-heb	flores200-devtest	show	17.4
20	slv-heb	ntrex128	show	15.4
21	ukr-heb	flores101-devtest	show	20.4
22	ukr-heb	flores200-devtest	show	19.8
23	ukr-heb	ntrex128	show	13.8
24	ukr-heb	tatoeba-test-v2020-07-28	show	43.1
25	ukr-heb	tatoeba-test-v2021-03-30	show	42.9
26	ukr-heb	tatoeba-test-v2021-08-07	show	43.2
			average	21.926