Authors
Adrian Mladenic,
Erik Novak,
Dunja Mladenic,
Publication date
Publisher
Total citations
Cited by
Description
This paper presents Slovenian commonsense description models based on the COMET framework for English. Inspired by MultiCOMETs approach to multilingual commonsense description, we finetune two Slovenian GPT-2 language models. Experimental evaluation based on several performance metrics shows comparable performance to the original COMET GPT-2 model for English.