monsoon-nlp/dv-wave

This is a second attempt at a Dhivehi language model trained with Google Research's ELECTRA and SimpleTransformers . The model was trained on @Sofwath's 307MB corpus of text: https://://://colab.google.com/drive/1ZJ3tttU9MwyWj6UtQ-8G7QJKTn-hG1uQ9v?usp=sharing. The model has a similar performance to mBERT on news classification task after finetuning for 3 epochs (52%) V2: fixed tokenizers do_lower_case=False and strip_accents=False to preserve vowel,