Repository logoRepository logo
 

NCHLT isiZulu word2vec-Skipgram embeddings

dc.contact.emailRoald.Eiselen@nwu.ac.za
dc.contact.nameRoald Eiselen
dc.contributor.authorRoald Eiselen
dc.contributor.otherRico Koen
dc.contributor.otherAlbertus Kruger
dc.contributor.otherJacques van Heerden
dc.date.accessioned2023-07-28T08:12:34Z
dc.date.accessioned2023-05-01
dc.date.available2023-07-28T08:12:34Z
dc.date.available2023-05-01
dc.date.issued2023-05-01
dc.descriptionStatic word embeddings for the Skipgram flavour of the word2vec (w2v) architecture (Mikolov et al., 2013). The embedding provides real-valued vector representations for isiZulu text.
dc.format.extentTraining data: Paragraphs: 816,776; Token count: 15,801,081; Vocab size: 101,924; Embedding dimensions: 600;
dc.format.size218.64MB (Zipped)
dc.identifier.urihttps://hdl.handle.net/20.500.12185/665
dc.language.isozu
dc.languagesisiZulu
dc.media.categoryWord embeddings
dc.media.typeText
dc.projectNCHLT Text IV
dc.publisherNorth-West University; Centre for Text Technology (CTexT)
dc.rights.licenseCreative Commons Attribution 4.0 International (CC-BY 4.0)
dc.software.requirementsPython
dc.sourceWeb
dc.sourceGovernment Documents
dc.titleNCHLT isiZulu word2vec-Skipgram embeddings
dc.typeModules

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
zu.w2v.model.skipgram.tar.gz
Size:
218.64 MB
Format:
Unknown data format