EvaLatin 2020 Models #evalatin20_models
EvaLatin 2020 Models are distributed under the CC BY-NC-SA licence. The models are based solely on EvaLatin 2020 treebanks, and additionally use multilingual BERT.
The models require UDPipe 2.
Download
The latest version 200831 of the EvaLatin 2020 models can be downloaded from LINDAT/CLARIN repository.
The models are also available in the REST service.
Acknowledgements #evalatin20_models_acknowledgements
This work was supported by the grant no. GX20-16819X of the Grant Agency of the Czech Republic, and has been using language resources stored and distributed by the LINDAT/CLARIAH-CZ project of the Ministry of Education, Youth and Sports of the Czech Republic (project LM2018101).
The models were trained on EvaLatin 2020 treebanks.
Finally, multilingual BERT is used to provide contextualized word embeddings.
Publications
- Milan Straka, Jana Straková (2020): UDPipe at EvaLatin 2020: Contextualized Embeddings and Treebank Embeddings. In: ArXiv.org Computing Research Repository, ISSN 2331-8422, 2006.03687
Model Performance
| Model | Dataset | UPOS | Lemma |
|---|---|---|---|
| latin-evalatin20-200830 | test classical | 96.73 | 96.39 |
| latin-evalatin20-200830 | test cross-genre | 90.47 | 86.89 |
| latin-evalatin20-200830 | test cross-time | 87.58 | 90.59 |