CroSloEngual BERT

CroSloEngual BERT is a trilingual model, using bert-base architecture, trained on Croatian, Slovenian, and English corpora. Focusing on three languages, the model performs better than multilingual BERT, while still offering an option for cross-lingual knowledge transfer, which a monolingual model wouldn't.

Evaluation is presented in our article:

@Inproceedings{ulcar-robnik2020finest,
author = "Ulčar, M. and Robnik-Šikonja, M.",
year = 2020,
title = "{FinEst BERT} and {CroSloEngual BERT}: less is more in multilingual models",
editor = "Sojka, P and Kopeček, I and Pala, K and Horák, A",
booktitle = "Text, Speech, and Dialogue {TSD 2020}",
series = "Lecture Notes in Computer Science",
volume = 12284,
publisher = "Springer",
url = "https://doi.org/10.1007/978-3-030-58323-1_11",
}

The preprint is available at arxiv.org/abs/2006.07890.

New

Select AutoNLP in the “Train” menu to fine-tune this model automatically.

Downloads last month
959
Hosted inference API
Fill-Mask
Mask token: [MASK]
This model can be loaded on the Inference API on-demand.