semantic-similarity text
Tatoeba
1,000 English-aligned sentence pairs for each language based on the Tatoeba corpus
Domains: Written
Languages: AfrikaansAmharicOld English (ca. 450-1100)ArabicAlgerian ArabicEgyptian ArabicAsturianAwadhiAzerbaijaniBelarusianBengaliBerber (Other)BosnianBretonBulgarianCatalanChavacanoCebuanoCzechChamorroMandarin ChineseCornishKashubianWelshDanishGermanLower SorbianKadazan DusunModern Greek (1453-)EnglishEsperantoEstonianBasqueFaroeseFinnishFrenchWestern FrisianScottish GaelicIrishGalicianSwiss GermanHebrewHindiCroatianUpper SorbianHungarianArmenianIdoInterlingueInterlingua (International Auxiliary Language Association)IndonesianIcelandicItalianJavaneseJapaneseKabyleGeorgianKazakhKhmerKoreanKurdishCoastal KadazanLatinLingua Franca NovaLithuanianStandard LatvianMalayalamMarathiNorth Moluccan MalayEastern MariMacedonianMongolianLow GermanDutchNorwegian NynorskNorwegian BokmålNovialOccitan (post 1500)Old RussianPampangaIranian PersianPiemontesePolishPortugueseRomanianRussianSlovakSlovenianSpanishAlbanianSerbianSwedishSwabianSwahili (individual language)TamilTatarTeluguTagalogThaiTurkmenTurkishTalossanUighurUkrainianUrduUzbekVietnameseWaray (Philippines)Wu ChineseXhosaYiddishYue ChineseStandard Malay
- Task type
- BitextMining
- Main metric
f1- Source dataset
- mteb/tatoeba-bitext-mining
- License
- cc-by-2.0
- Dates
- 2006-01-01 → 2021-12-31
- Annotations
- human-annotated
- Sample creation
- found
Models scored …