BanglaBERT: Language Model Pretraining and Benchmarks for Low-Resource Language Understanding Evaluation in Bangla

In this work, we introduce <PRE_TAG>BanglaBERT</POST_TAG>, a BERT-based Natural Language Understanding (NLU) model pretrained in Bangla, a widely spoken yet low-resource language in the NLP literature. To pretrain <PRE_TAG>BanglaBERT</POST_TAG>, we collect 27.5 GB of Bangla pretraining data (dubbed `<PRE_TAG>Bangla2B+</POST_TAG>') by crawling 110 popular Bangla sites. We introduce two downstream task datasets on natural language inference and question answering and benchmark on four diverse NLU tasks covering text classification, sequence labeling, and span prediction. In the process, we bring them under the first-ever Bangla Language Understanding Benchmark (BLUB). <PRE_TAG>BanglaBERT</POST_TAG> achieves state-of-the-art results outperforming multilingual and monolingual models. We are making the models, datasets, and a leaderboard publicly available at https://github.com/csebuetnlp/banglabert to advance Bangla NLP.