Forums

> Training corpus released

Dear participants,

The training data (with labels) and the development data (without labels) are published in the "Files" section.
During this phase, you can use the training data to create your systems, and submit your predictions for the development data.

Regards,
Luis

Posted by: luischir @ March 22, 2023, 8:52 p.m.

---

Hi,

I did not see a definition whether the task is "open" or "closed"? Can we use existing Spanish and Guarani monolingual corpora or pre-trained models (like BERT or HeLI-OTS)?

Cheers,
Tommi

Posted by: tsjauhia @ March 25, 2023, 12:10 p.m.

---

Hi,

You can use any external data or pretrained models you want, except the dev and test splits of the Jojajovai corpus.

Regards,
Luis

Posted by: luischir @ March 25, 2023, 6:56 p.m.