Fine Tune For Custom Hindi Dataset

#2
by D3v - opened

Hey Guys, Can you tell me in which schema i have to annotate my own Hindi corpus and create my dataset then fine tune on this model ?

AI4Bharat org

Hi, the data used for training the model follows BIO notation. The model is already fine-tuned for Named Entity Recognition task on 11 Indic languages. You could further fine-tune the model(domain adaptation) on your Hindi corpus.

Actually i want to create NER notation from my own domain specific Hindi Corpus , I need to tag words and have like more than 15 labels , But some word are splitted that time model is unable to recognise . Can you tell me how i can annotate in BIO notation having perfect labels.

D3v changed discussion status to closed

Sign up or log in to comment