Pre-training Data Quality and Quantity for a Low-Resource Language: New Corpus and BERT Models for Maltese

Open in new window