Pretrained models for tagging and parsing Swedish are released

Den här sidan är inte översatt till svenska. Innehållet visas därför på engelska.
Språkbanken releases pretrained models for POS tagging and dependency parsing of Swedish texts, as well as a list of available Swedish embeddings

We have added several pretrained Swedish models to our downloadable resources. If you want to add morphological or syntactic annotation to a Swedish text, but do not have time to a perform a full-scale investigation in order to find out which tools yield best results and what they have to be trained on, do not worry: we have done the work for you. With our models, state-of-the-art tagging and parsing should be easy.

If you need POS and MSD tags (SUC style), choose one of the six POS models (each with its own advantages). If you need dependency annotation (Mamba-Dep style), choose one of the two syntactic models. If you are looking for Swedish embeddings (which may serve many useful purposes), have a look at the list we compiled.

Using the models does not require deep knowledge of natural language processing or advanced programming skills.

More to come.