CroSloEngual BERT 1.1

  Avtorji

Ulčar, Matej and Robnik-Šikonja, Marko

  Identifikator vnosa

http://hdl.handle.net/11356/1330

 URL projekta

http://embeddia.eu

 Dokumentirano v

https://arxiv.org/abs/2006.07890

 Datum objave

2020-07-09

 Vrsta

toolService

 Jezik(i)

Croatian , English , Slovenian

 Opis

Trilingual BERT (Bidirectional Encoder Representations from Transformers) model, trained on Croatian, Slovenian, and English data. State of the art tool representing words/tokens as contextually dependent word embeddings, used for various NLP classification tasks by finetuning the model end-to-end. CroSloEngual BERT are neural network weights and configuration files in pytorch format (i.e. to be used with pytorch library). Changes in version 1.1: fixed vocab.txt file, as previous verson had an error causing very bad results during fine-tuning and/or evaluation.

 Izdajatelj

Faculty of Computer and Information Science, University of Ljubljana

 Zahvala

European Union EC/H2020/825153 "EMBEDDIA - Cross-Lingual Embeddings for Less-Represented Languages in European News Media"

 Ključne besede

word embeddings multilingual contextual embeddings BERT language model

 Zbirke

CLARIN.SI data & tools

 Druge različice

Seznam vseh različic  

Prikaži polni zapis vnosa