README.md What to do: tokenisation - auto tokeniser data augmentation / preprocessing - synonym (wordnet) (azhara) - remove all caps - add text to samples - back translation (ella) - feature space synonym replacemnet (emily) one other newer model than roberta sam