forked from Helsinki-NLP/Tatoeba-Challenge
-
Notifications
You must be signed in to change notification settings - Fork 0
/
opus-2020-07-27.yml
53 lines (53 loc) · 1.14 KB
/
opus-2020-07-27.yml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
release: sem-sem/opus-2020-07-27.zip
release-date: 2020-07-27
dataset-name: opus
modeltype: transformer
pre-processing: normalization + SentencePiece (spm32k,spm32k)
subwords:
- source: spm32k
- target: spm32k
subword-models:
- source: source.spm
- target: target.spm
source-languages:
- apc
- ara
- arq
- arz
- heb
- mlt
target-languages:
- apc
- ara
- arq
- arz
- heb
- mlt
use-target-labels:
- >>apc<<
- >>ara<<
- >>arq<<
- >>arz<<
- >>heb<<
- >>mlt<<
test-data:
- Tatoeba-test.ara-ara: 16/48
- Tatoeba-test.ara-heb: 1208/5440
- Tatoeba-test.ara-mlt: 28/59
- Tatoeba-test.heb-ara: 1208/5301
- Tatoeba-test.mlt-ara: 28/67
- Tatoeba-test.multi-multi: 2504/10963
BLEU-scores:
- Tatoeba-test.ara-ara: 4.2
- Tatoeba-test.ara-heb: 34.0
- Tatoeba-test.ara-mlt: 16.6
- Tatoeba-test.heb-ara: 18.8
- Tatoeba-test.mlt-ara: 20.7
- Tatoeba-test.multi-multi: 27.1
chr-F-scores:
- Tatoeba-test.ara-ara: 0.200
- Tatoeba-test.ara-heb: 0.542
- Tatoeba-test.ara-mlt: 0.513
- Tatoeba-test.heb-ara: 0.477
- Tatoeba-test.mlt-ara: 0.388
- Tatoeba-test.multi-multi: 0.507