paraphrase_set_id
stringlengths
1
6
sentence_id
stringlengths
1
7
paraphrase
stringlengths
1
37.4k
lists
sequencelengths
1
18
tags
sequencelengths
1
9
language
stringclasses
73 values
0
5494428
Ek drink.
[ "" ]
[ "" ]
af
0
5875986
Ek is dors.
[ "" ]
[ "" ]
af
0
7986816
Ek drink water.
[ "" ]
[ "" ]
af
40
5458722
Dis binnekort lente.
[ "" ]
[ "" ]
af
40
5458724
Dit is lente.
[ "" ]
[ "" ]
af
40
5458725
Die lente het teruggekom.
[ "" ]
[ "" ]
af
40
5458726
Die lente kom nader.
[ "" ]
[ "" ]
af
56
846136
Ek soek werk.
[ "" ]
[ "" ]
af
56
1597519
Ek werk.
[ "" ]
[ "" ]
af
56
5443512
Ek studeer.
[ "" ]
[ "" ]
af
56
5997523
Ek is tans besig.
[ "" ]
[ "" ]
af
88
675962
Dit is my vriend.
[ "" ]
[ "" ]
af
88
753260
Dis my vriend.
[ "" ]
[ "" ]
af
88
1357207
Dis my dogter.
[ "" ]
[ "" ]
af
88
7986875
Sy is my meisie.
[ "" ]
[ "" ]
af
92
1359766
My broer leef in Tokio.
[ "" ]
[ "" ]
af
92
1598611
My broer woon in Tokio.
[ "" ]
[ "" ]
af
94
1386997
Die Titanic het gesink op haar eerste vaart. Sy was 'n groot skip.
[ "" ]
[ "" ]
af
94
1598593
Die Titanic het op sy eerste vaart gesink. Dit was 'n groot skip.
[ "" ]
[ "" ]
af
98
1422322
Die landskap is awemrowend.
[ "" ]
[ "" ]
af
98
1598552
Die landskap is asemrowend.
[ "" ]
[ "" ]
af
106
1422336
Ons sal jou roep wanneer jou tafel gereed is.
[ "" ]
[ "" ]
af
106
1422339
Ons sal u skakel wanneer u tafel gereed is.
[ "" ]
[ "" ]
af
118
1424410
Hoe lank het jy daar gebly?
[ "" ]
[ "" ]
af
118
1424411
Hoe lank het u daar gebly?
[ "" ]
[ "" ]
af
155
1597561
Wag totdat die verkeerslig groen word.
[ "" ]
[ "" ]
af
155
5757201
Wag totdat die lig groen is.
[ "" ]
[ "" ]
af
167
983996
Hierdie hond is van my.
[ "" ]
[ "" ]
af
167
1598592
Hierdie hond is myne.
[ "" ]
[ "" ]
af
169
385198
Kan jy jou kar beweeg asseblief?
[ "" ]
[ "" ]
af
169
1598640
Kan jy asseblief jou kar skuif.
[ "" ]
[ "" ]
af
169
2779458
Sal jy asseblief jou motor skuif?
[ "" ]
[ "" ]
af
170
384993
Hul het hul seun John benoem.
[ "" ]
[ "" ]
af
170
1598642
Hulle het hulle seun John genoem.
[ "" ]
[ "" ]
af
190
1611267
Van watter sportsoort hou jy die meeste?
[ "" ]
[ "" ]
af
190
2788467
Van watse sport hou jy die meeste van?
[ "" ]
[ "" ]
af
199
1706352
Ons het met die hysbak op en af gery.
[ "" ]
[ "" ]
af
199
1857693
Ons het op en af met die hysbak gery.
[ "" ]
[ "" ]
af
214
1735096
Jy moet my gewis more oggend bel.
[ "" ]
[ "" ]
af
214
1735097
Maak seker dat jy my more oggend bel.
[ "" ]
[ "" ]
af
295
1759574
Tom het gedink dit was onregverdig.
[ "" ]
[ "" ]
af
295
1759576
Tom dog dat dit onregverdig was.
[ "" ]
[ "" ]
af
310
1764509
Hulle het wyn.
[ "" ]
[ "" ]
af
310
1764513
Sy het wyn.
[ "" ]
[ "" ]
af
312
1764511
Jy het wyn.
[ "" ]
[ "" ]
af
312
1764512
Julle het wyn.
[ "" ]
[ "" ]
af
336
1766555
Sy is jonk.
[ "" ]
[ "" ]
af
336
1766557
Hy is jonk.
[ "" ]
[ "" ]
af
336
4874643
Sy was jonk.
[ "" ]
[ "" ]
af
340
1766992
Hy kan 'n bietjie Engels praat.
[ "" ]
[ "" ]
af
340
1766993
Hy kan nie veel Engels praat nie.
[ "" ]
[ "" ]
af
353
1769347
Het u 'n selfoon?
[ "" ]
[ "" ]
af
353
1769348
Het julle 'n selfoon?
[ "" ]
[ "" ]
af
353
1769355
Het jy 'n selfoon?
[ "" ]
[ "" ]
af
379
1775789
Wat dink jy moet gedoen word daaroor?
[ "" ]
[ "" ]
af
379
1775790
Volgens jou wat moet daaroor gedoen word?
[ "" ]
[ "" ]
af
391
1783258
Water is belangrik.
[ "" ]
[ "" ]
af
391
5945834
Die meer is groot.
[ "" ]
[ "" ]
af
396
1784091
Waar is my pa?
[ "" ]
[ "" ]
af
396
5489351
Waar is pa?
[ "" ]
[ "" ]
af
397
1784122
Ons sal jou begrawe.
[ "" ]
[ "" ]
af
397
1784124
Ons sal julle begrawe.
[ "" ]
[ "" ]
af
423
1802672
Hou jou vriend van tee?
[ "" ]
[ "" ]
af
423
1802674
Hou jou vriendin van tee?
[ "" ]
[ "" ]
af
488
1823593
Watter soort boek wil jy he?
[ "" ]
[ "" ]
af
488
2788468
Watse boek wil jy hê?
[ "" ]
[ "" ]
af
525
1838401
Die oomblik wat ek haar gesien het, het ek haar herken.
[ "" ]
[ "" ]
af
525
3437716
Ek het haar herken toe ek haar skaars gesien het.
[ "" ]
[ "" ]
af
563
1854974
Ek werk by hierdie maatskappy.
[ "" ]
[ "" ]
af
563
5491767
Ek werk vir hierdie onderneming.
[ "" ]
[ "" ]
af
576
1857695
Het jy 'n vuurhoutjie?
[ "" ]
[ "" ]
af
576
1857696
Het u vuurhoutjies?
[ "" ]
[ "" ]
af
582
384916
Ek kan nie nou aan sy naam dink nie.
[ "" ]
[ "" ]
af
582
1860077
Vir 'n oomblik het ek vergeet wat haar naam was.
[ "" ]
[ "" ]
af
610
1915379
Sy is bly oor haar nuwe rok.
[ "" ]
[ "" ]
af
610
6607409
Sy was bly met haar nuwe rok.
[ "" ]
[ "@needs native check" ]
af
616
1918407
Hy het gepraat.
[ "" ]
[ "" ]
af
616
1918408
Hy was besig om te praat.
[ "" ]
[ "" ]
af
659
1974528
Waar bly jou oupa?
[ "" ]
[ "" ]
af
659
5897622
Waar woon jou oupa?
[ "" ]
[ "" ]
af
688
387068
Ek het laas nag hier aangekom.
[ "" ]
[ "" ]
af
688
2019902
Ek het gister gekom.
[ "" ]
[ "" ]
af
694
2045062
My vriend het my 'n brief geskryf waarin hy my gevra het of dit goed gaan met my.
[ "" ]
[ "" ]
af
694
4263597
My vriend het vir my 'n brief geskryf om uit te vind hoe ek was.
[ "" ]
[ "" ]
af
714
2060247
Skakel af die lig.
[ "" ]
[ "" ]
af
714
2364248
Kan jy die ligte afskakel?
[ "" ]
[ "" ]
af
755
2345891
Is jou ma tuis?
[ "" ]
[ "" ]
af
755
5540661
Is jou ma hier?
[ "" ]
[ "" ]
af
946
2779400
Ek het vir Ann 'n pop gemaak.
[ "" ]
[ "" ]
af
946
2779403
Ek het 'n pop vir Ann gemaak.
[ "" ]
[ "" ]
af
948
2779407
Daar was 'n bus skedule op die muur.
[ "" ]
[ "" ]
af
948
2779409
Op die muur was daar 'n bus skedule.
[ "" ]
[ "" ]
af
1093
2783911
Ek sal dit terug gee.
[ "" ]
[ "" ]
af
1093
2783919
Ek gee dit terug.
[ "" ]
[ "" ]
af
1107
2783932
Ek is al gewoond daaraan teen nou.
[ "" ]
[ "" ]
af
1107
2970746
Ek is nou al gewoond daaraan.
[ "" ]
[ "" ]
af
1110
2783937
Ek sou nie so seker wees nie.
[ "" ]
[ "" ]
af
1110
2784157
Ek sou nie te seker wees daaroor nie.
[ "" ]
[ "" ]
af
1113
2783942
Ek sal die volgende bus vang.
[ "" ]
[ "" ]
af
1113
2784140
Ek gaan op die volgende bus wees.
[ "" ]
[ "" ]
af

Dataset Card for TaPaCo Corpus

Dataset Summary

A freely available paraphrase corpus for 73 languages extracted from the Tatoeba database. Tatoeba is a crowdsourcing project mainly geared towards language learners. Its aim is to provide example sentences and translations for particular linguistic constructions and words. The paraphrase corpus is created by populating a graph with Tatoeba sentences and equivalence links between sentences “meaning the same thing”. This graph is then traversed to extract sets of paraphrases. Several language-independent filters and pruning steps are applied to remove uninteresting sentences. A manual evaluation performed on three languages shows that between half and three quarters of inferred paraphrases are correct and that most remaining ones are either correct but trivial, or near-paraphrases that neutralize a morphological distinction. The corpus contains a total of 1.9 million sentences, with 200 – 250 000 sentences per language. It covers a range of languages for which, to our knowledge, no other paraphrase dataset exists.

Supported Tasks and Leaderboards

Paraphrase detection and generation have become popular tasks in NLP and are increasingly integrated into a wide variety of common downstream tasks such as machine translation , information retrieval, question answering, and semantic parsing. Most of the existing datasets cover only a single language – in most cases English – or a small number of languages. Furthermore, some paraphrase datasets focus on lexical and phrasal rather than sentential paraphrases, while others are created (semi -)automatically using machine translation.

The number of sentences per language ranges from 200 to 250 000, which makes the dataset more suitable for fine-tuning and evaluation purposes than for training. It is well-suited for multi-reference evaluation of paraphrase generation models, as there is generally not a single correct way of paraphrasing a given input sentence.

Languages

The dataset contains paraphrases in Afrikaans, Arabic, Azerbaijani, Belarusian, Berber languages, Bulgarian, Bengali , Breton, Catalan; Valencian, Chavacano, Mandarin, Czech, Danish, German, Greek, Modern (1453-), English, Esperanto , Spanish; Castilian, Estonian, Basque, Finnish, French, Galician, Gronings, Hebrew, Hindi, Croatian, Hungarian , Armenian, Interlingua (International Auxiliary Language Association), Indonesian, Interlingue; Occidental, Ido , Icelandic, Italian, Japanese, Lojban, Kabyle, Korean, Cornish, Latin, Lingua Franca Nova\t, Lithuanian, Macedonian , Marathi, Bokmål, Norwegian; Norwegian Bokmål, Low German; Low Saxon; German, Low; Saxon, Low, Dutch; Flemish, ]Old Russian, Turkish, Ottoman (1500-1928), Iranian Persian, Polish, Portuguese, Rundi, Romanian; Moldavian; Moldovan, Russian, Slovenian, Serbian, Swedish, Turkmen, Tagalog, Klingon; tlhIngan-Hol, Toki Pona, Turkish, Tatar, Uighur; Uyghur, Ukrainian, Urdu, Vietnamese, Volapük, Waray, Wu Chinese and Yue Chinese

Dataset Structure

Data Instances

Each data instance corresponds to a paraphrase, e.g.:

{ 
    'paraphrase_set_id': '1483',  
    'sentence_id': '5778896',
    'paraphrase': 'Ɣremt adlis-a.', 
    'lists': ['7546'], 
    'tags': [''],
    'language': 'ber'
}

Data Fields

Each dialogue instance has the following fields:

  • paraphrase_set_id: a running number that groups together all sentences that are considered paraphrases of each other
  • sentence_id: OPUS sentence id
  • paraphrase: Sentential paraphrase in a given language for a given paraphrase_set_id
  • lists: Contributors can add sentences to list in order to specify the original source of the data
  • tags: Indicates morphological or phonological properties of the sentence when available
  • language: Language identifier, one of the 73 languages that belong to this dataset.

Data Splits

The dataset is having a single train split, contains a total of 1.9 million sentences, with 200 – 250 000 sentences per language

Dataset Creation

Curation Rationale

[More Information Needed]

Source Data

Initial Data Collection and Normalization

[More Information Needed]

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

[More Information Needed]

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

[More Information Needed]

Licensing Information

Creative Commons Attribution 2.0 Generic

Citation Information

@dataset{scherrer_yves_2020_3707949,
  author       = {Scherrer, Yves},
  title        = {{TaPaCo: A Corpus of Sentential Paraphrases for 73 Languages}},
  month        = mar,
  year         = 2020,
  publisher    = {Zenodo},
  version      = {1.0},
  doi          = {10.5281/zenodo.3707949},
  url          = {https://doi.org/10.5281/zenodo.3707949}
}

Contributions

Thanks to @pacman100 for adding this dataset.

Downloads last month
3,384

Models trained or fine-tuned on community-datasets/tapaco