Skip to main content


Talbanken is a Swedish treebank.

Talbanken is a widely used Swedish treebank, read more about its history and different versions here. This version was created by Joakim Nivre et al. at Uppsala University as part of The Swedish Treebank (STB). It is now being distributed by Språkbanken Text, and for reproducibility’s sake we keep it in exactly the same form as it was when we received it, without any corrections. This version is not indexed by our search engine Korp.


The following layers of annotation were added (or corrected) manually and can be considered gold data: tokenization, sentence segmentation, POS, MSD, dependency syntax (deprel and dephead).

Tokenization, sentence segmentation, POS and MSD follow the SUC format, syntactic annotation follows the Mamba-Dep format, a conversion of the MAMBA format used in the original Talbanken76 to dependency grammar.

Read more about these annotation layers in the documentation or at Joakim Nivre's page: tokenization and sentence segmentation, POS and MSD, dependency syntax.

All other annotation, including lemmatization, was created automatically by Sparv.

Formats and splits

TalbankenSTB is available in the CONLL and XML formats. A train-test split is available.

File Size Modified Licence (zip)
2.6 MB 2020-08-11 CC BY 4.0
TalbankenSTB_README.txt (txt)
1.05 KB 2020-08-11 CC BY 4.0
attribution (zip)
62.23 KB 2020-08-11 CC BY 4.0
attribution (zip)
2.6 MB 2020-08-11 CC BY 4.0
attribution (zip)
2.95 MB 2020-08-11 CC BY 4.0


  • Corpus
  • Training and evaluation data




Tokens: 96,346
Sentences: 6,160