roberta-base model is fine tuned on Kaggle two datasets namely "Sentiment Analysis Dataset" & Pre-processed Twitter tweets.

Respective labels and thier sentiment