A RNN based model for predicitng the sentiment of sentences in the Stanford Sentiment Tree Binary Dataset (available at https://nlp.stanford.edu/sentiment/). The model predicts whether a sentence has positive or negative sentiment.
This model was built using the fastai.text open source library (Jeremy Howard et al.). The model achieved a validation accuracy of ~93%.
If you would like to see this model deployed into a Flask API please see my other repository: https://github.com/brandonfonseca/ml-flask-api