Since this can be a classification downside, we’ll use a "many to one" RNN.
Later in this put up, we’ll construct a "many to one" RNN from scratch to perform fundamental Sentiment Evaluation.