
Institute · Stockholm, Sweden
KBLabNational Library of Sweden
Data lab inside the National Library of Sweden that trains open Swedish language and speech models, including KB-BERT and KB-Whisper, on the library's collections.
Stockholm, Sweden
- Type
- Institute
- Based in
- Stockholm, Sweden
- Key papers
- 3
- Featured projects
- 4
About
KBLab is a research department within the National Library of Sweden, set up in 2019 as a national research infrastructure for digital humanities and social science. It gives researchers access to the library's collections in structured form. The lab is directed by Love Börjeson.
Using the legal-deposit and digitized holdings, KBLab trains Swedish models. Its 2020 paper by Martin Malmsten, Love Börjeson and Chris Haffenden introduced KB-BERT. The lab later released KB-Whisper, speech recognition models trained on more than 50,000 hours of Swedish speech, which it reports outperform OpenAI's Whisper at each size on Swedish.
Models are published on Hugging Face, with the large KB-Whisper model under the Apache 2.0 license.
Mission and ambitionsKBLab aims to make the National Library's collections usable for research through language and speech models and open tools, supporting digital humanities and social science.
Known for
- 01
KB-BERT Swedish language model
- 02
KB-Whisper Swedish speech recognition
- 03
Models trained on national library collections
Milestones
- 2019
KBLab established
The National Library set up an integrated data lab for research.
- 2020
KB-BERT
Published the paper on building a Swedish BERT model from the library's collections.
- 2025
KB-Whisper
Released Swedish Whisper models trained on over 50,000 hours of speech, in sizes from tiny to large.
Open source
Model: 728,037 downloads, 26 likes.
Model: 38,229 downloads, 37 likes.
- KBLab/rixvox-v2dataset
Dataset: 5,545 downloads, 14 likes.
Space: 4 likes on the Hub.
Research papers
Notable releases
Openness
Models and datasets are public on Hugging Face; check each card for licence terms.
People
- Love BörjesonDirector of KBLab
- Martin MalmstenAuthor of the KB-BERT paper
- Chris HaffendenAuthor of the KB-BERT paper
Listed from the lab's own pages and public profiles.
Contact and community
Contact via the KBLab page on the National Library website.
Sources
Compiled from the lab's public pages, Hugging Face and GitHub. Last checked 2026-11-17. Something wrong or out of date? Tell us on Discord.



