Partners

A project of the complexity and magnitude of Aina requires the collaboration of many entities and people. We appreciate the collaboration that makes the Aina project possible.

We also work very closely with the “ILENIA Project“ (Impulso de las Lenguas en Inteligencia Artificial), a project framed within the PERTE de Nueva Economia de la Lengua. The other project partners are HiTZ, USC and CENID

It also collaborates with similar projects such as:

For the generation of the Aina Kit datasets, we have worked with many organizations to obtain the volume of data necessary to be able to train the Aina models.

Government entities

Media

Academic and book repositories

Other collaborations

Collaboration with the Parla’m project promoted by Òmnium Cultural for the collection of textual and voice data.
To date, the following entities have signed the Content Transfer Agreement
Llenguaferits (text and audio)
BIFIDUS PRODUCCIONS SL (text and audio)
Unió de Pagesos (text and audio)
Grup Món (text)
Núvol (text)
Betevé (audio)

Reuse of pre-existing datasets

Pre-existing data sets have been used, such as:

Open data accessible thanks to repositories such as

Companies with which the BSC has collaborated in the curation, creation or translation of instructions, labelling, etc.
Process Talks
M47Labs
Open.Voice.OS
TALP – UPC
UB Speech

Collaborations with other open LLM models such as

Using infrastructure and support from

Integration of Aina models in commercial products

Integration with open source environments

Major source code repositories