Home / Query / WordAlign / Wiki     [books] [DGT] [DOGC] [ECB] [EMEA] [EUbooks] [EU] [Europarl] [GNOME] [GlobalVoices] [hren] [JRC] [KDE4/doc] [MBS] [MultiUN] [NCv9/v11] [OO/OO3] [subs/16/18] [ParCor] [PHP] [SETIMES] [SPC] [Tatoeba] [TEP] [TedTalks] [TED] [Tanzil] [Ubuntu] [UN] [WikiSource] [Wikipedia] [WMT]

Wikipedia

This is a corpus of parallel sentences extracted from Wikipedia by Krzysztof Wołk and Krzysztof Marasek. Please cite the following publication if you use the data: Krzysztof Wołk and Krzysztof Marasek: Building Subject-aligned Comparable Corpora and Mining it for Truly Parallel Sentence Pairs., Procedia Technology, 18, Elsevier, p.126-132, 2014

20 languages, 36 bitexts
total number of files: 114
total number of tokens: 610.13M
total number of sentence fragments: 25.90M

Please cite the following article if you use any part of the corpus in your own work:
Jörg Tiedemann, 2012, Parallel Data, Tools and Interfaces in OPUS. In Proceedings of the 8th International Conference on Language Resources and Evaluation (LREC 2012)

Download

Complete download of aligned documents (in XML): Wikipedia1.0.tar.gz (7.8G )

Bottom-left triangle: download files
  • ces = sentence alignments in XCES format
  • leftmost column language IDs = tokenized corpus files in XML
  • TMX and plain text files (Moses): see "Statistics" below
  • lower row language IDs = parsed corpus files (if they exist)
Upper-right triangle: sample files
  • view = bilingual XML file samples
  • upper row language IDs = monolingual XML file samples
  • rightmost column language IDs = untokenized corpus files

ar bg cs de el en es fa fr he hu it nl pl pt ro ru sl tr vi
ar viewviewar
bg viewbg
cs viewviewcs
de viewviewde
el viewviewel
en ces ces ces ces ces viewviewviewviewviewviewviewviewviewviewviewviewviewviewen
es ces viewes
fa ces viewfa
fr ces viewfr
he ces viewhe
hu ces viewhu
it ces viewit
nl ces viewnl
pl ces ces ces ces ces ces ces ces ces ces ces ces viewviewviewviewviewviewpl
pt ces ces pt
ro ces ces ro
ru ces ces ru
sl ces ces sl
tr ces ces tr
vi ces ces vi
ar bg cs de el en es fa fr he hu it nl pl pt ro ru sl tr vi

Statistics and TMX/Moses Downloads

Number of files, tokens, and sentences per language (including non-parallel ones if they exist)
Number of sentence alignment units per language pair

Upper-right triangle: download translation memory files (TMX)
Bottom-left triangle: download plain text files (MOSES/GIZA++)
Language ID's, first row: monolingual plain text files (tokenized)
Language ID's, first column: monolingual plain text files (untokenized)

language files tokens sentencesar bg cs de el en es fa fr he hu it nl pl pt ro ru sl tr vi
ar 3 24.9M 1.0M0.1M 0.8M
bg 1 1.4M 79.8k77.1k
cs 2 1.7M 90.2k25.4k 61.5k
de 7 45.5M 2.6M2.4M 0.2M
el 2 2.3M 0.1M0.1M 12.1k
en 39 275.8M 10.7M0.2M 79.8k 27.7k 2.5M 0.1M 1.7M 95.7k 0.8M 0.1M 56.1k 0.9M 0.8M 0.2M 1.7M 0.4M 0.5M 81.5k 0.2M 36.3k
es 6 44.5M 2.0M1.8M 0.1M
fa 2 2.7M 0.1M97.1k 6.0k
fr 3 18.8M 0.9M0.8M 49.8k
he 2 3.0M 0.1M0.1M 9.9k
hu 2 1.9M 0.1M61.5k 40.2k
it 4 27.2M 1.2M1.0M 0.2M
nl 3 21.8M 1.0M0.8M 0.2M
pl 19 52.4M 2.4M0.8M 62.5k 0.2M 12.2k 0.2M 0.2M 6.1k 51.7k 10.0k 41.1k 0.2M 0.2M 0.2M 6.6k 0.2M 17.0k 15.9k 89.6k
pt 6 50.2M 2.0M1.8M 0.2M
ro 3 11.1M 0.4M0.4M 6.7k
ru 3 14.5M 0.7M0.6M 0.2M
sl 2 3.0M 0.2M0.1M 17.3k
tr 3 4.7M 0.2M0.2M 16.0k
vi 2 2.7M 0.1M58.1k 90.4k

Note that TMX files only contain unique translation units and, therefore, the number of aligned units is smaller than for the distributions in Moses and XML format. Moses downloads include all non-empty alignment units including duplicates. Token counts for each language also include duplicate sentences and documents.