Italian and French BERT: how to cite / acknowledge authorship

[ivandre]

We are using both Italian and French DBMDZ BERT in a multilingual research project, and I would like to acknowledge its authorship. Is there any publication to be cited? Website project? Or at least the authors' names? Many thanks.

[stefan-it]

Hi @ivandre ,

thanks for your interest in our models and using them 🤗

At the moment, there's no offical paper out there that can be cited. Other papers, that are using our models, just cite the repository url - I think that would be fine :) But I'm currently thinking about generating a Zenodo/DOI for that repository here, so it can be cited like the BERTurk models!

Btw: for Italian I will release an ELECTRA model quite soon (checkpoint selection and evaluations are done).

For French there are two exiting news: ELECTRA model for Europeana is on its way (already trained) and CamemELECTRA (same training data and vocab as CamemBERT) is also trained. So stay tuned!

[ivandre]

Excellent! Thank you very much indeed for your response and the great news. best regards Ivandré. Em seg., 2 de nov. de 2020 às 06:58, Stefan Schweter < [email protected]> escreveu:

…

Hi @ivandre <https://github.com/ivandre> , thanks for your interest in our models and using them 🤗 At the moment, there's no offical paper out there that can be cited. Other papers <https://github.com/dbmdz/berts/blob/main/papers.md>, that are using our models, just cite the repository url - I think that would be fine :) But I'm currently thinking about generating a Zenodo/DOI for that repository here, so it can be cited like the BERTurk <https://github.com/stefan-it/turkish-bert> models! Btw: for Italian I will release an ELECTRA model quite soon (checkpoint selection and evaluations are done). For French there are two exiting news: ELECTRA model for Europeana is on its way (already trained) and CamemELECTRA (same training data and vocab as CamemBERT) is also trained. So stay tuned! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#23 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALAU7FGVQ7QPJGEYWI4ADMDSNZ7F3ANCNFSM4TDWNFVA> .

[stefan-it]

@ivandre The Italian XXL ELECTRA model is now available on the Hugging Face model hub (model card is coming soon):

• Discriminator model
• Generator model

Repository for more information (like evaluations on NER/PoS tagging) is also available here.