The data preparation is done and now we take the produced matrices X_train and Y_train and use them for training a model. But of course, we need to create the model first. In this chapter, we will use a recurrent neural network with LSTM cells (Long Short-Term Memory). LSTM cells are better than plain RNN cells because they can capture long-term dependencies better.
A great resource to learn more about LSTMs is the blog post "Understanding LSTM Networks" by Christopher Olah, which is available at https://colah.github.io/posts/2015-08-Understanding-LSTMs/. In this chapter, we do not go into theoretical details about how LSTM and RNN work and only look at using them in TensorFow.
Let us start with defining ...
