Top 1 text-corpus open source projects

Polish RoBERTA model trained on Polish literature, Wikipedia, and Oscar. The major assumption is that quality text will give a good model.