NE-BERT: A Multilingual Language Model for Nine Northeast Indian Languages
NE-BERT, a domain-specific multilingual encoder model trained on approximately 8.3 million sentences spanning 9 Northeast Indian languages and 2 anchor languages, addresses critical vocabulary fragmentation issues in extremely low-resource languages such as Pnar and Kokborok through aggressive upsampling strategies.
Badal Nyalang
· Proceedings of the Second Wo... · 2 citations