[trying to get around with formatting issues for the digest version of corpora list]

[sending again: many apologies for repetition]

Good morning,

We are pleased to announce the release of Albertina PT-*

This is the first large language model specifically for Portuguese, covering both variants PT-PT and PT-BR, publicly available and open source.

With its 900 million parameters in this first version, its sets new state of the art for models specifically for Portuguese that are publicly available and open.

It was developed at the University of Lisbon together with colleagues from the University of Porto, and can be obtained here: https://huggingface.co/models?other=albertina-pt*

Its development is documented here: https://arxiv.org/abs/2305.06721

Best regards,

On behalf of Albertina's team