1.5-Pints Technical Report: Pretraining in Days, Not Months -- Your Language Model Thrives on Quality Data | Read Paper on Bytez