How the GPT2, BERT, and BART models were pre-trained on Vietnamese text, with the hyperparameters and the results each one reached.
