I would like to build a recurrent neural network (RNN) autoencoder, specifically using SimpleRNN and LSTMs, that shall mirror the input timeseries to the output layer. I have ti