est-roberta-hist-ner-for-tccp

Model description

est-roberta-hist-ner-for-tccp is an Est-RoBERTa based model fine-tuned for named entity recognition (NER) in the historical meeting protocols of the Tartu City Council from 1918โ€“1940, written in Estonian. Model training and evaluation experiments have been conducted by Sofia Kriuchkova, and the code is available in this repository. The following types of entities are recognized: organization names (ORG), person names (PER), locations (LOC), money sums (MONEY), person's occupations or roles (POSITION) and names of laws and regulations (LAW).

The model building work has been carried out under the project "Information extraction by the example of protocols of historical institutions (1880โ€“1940)" (EKKD-TA10), which is funded by the National Program "Estonian Language and Culture in the Digital Age".

Citation

If you use this model in your work, please cite us as follows:

@article{Orasmaa_Muischnek_Kriuchkova_2026, 
    author={Orasmaa, Siim and Muischnek, Kadri and Kriuchkova, Sofia}, 
    title={Named Entity Recognition in the Historical Meeting Protocols of the Tartu City Council}, 
    journal={Digital Humanities in the Nordic and Baltic Countries Publications}, 
    volume={8}, 
    DOI={10.5617/dhnbpub.13204}, 
    number={1}, 
    year={2026}, 
    month={Mar.}, 
    url={https://journals.uio.no/dhnbpub/article/view/13204} 
}
Downloads last month
31
Safetensors
Model size
0.1B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for tartuNLP/est-roberta-hist-ner-for-tccp

Finetuned
(4)
this model