This is a simple web app which predicts the sentiment of a given sentence/paragraph. It uses the LSTM model which was trianed on the sentiment-140 dataset. This dataset contain 1.6 million tweets. The link to the same dataset is here.
The link to the saved LSTM model file is here.