This is a sample reference model trained only on the Oscar Corpus for a day on a TPU v3-8. Pretrained model on Norwegian language using a causal language modeling (CLM) objective.
