Python implementation of one layer RNNs

Tensorflow has implemented many commonly-used RNN layers, including basic RNN, GRU, LSTM, etc. To construct a multi-layer RNN in Tensorflow, one simply stacks individual layers by sequentially adding each layer with specified input, output, and hidden size.

In this video, we show a minimal example of building a multi-layer RNN with/without intermediate output and reading the model summary.