This is from the RNN lesson from the Deep Learning Nanodegree at Udacity.
We'll build a character-wise RNN trained on Anna Karenina. It'll be able to generate new text based on the text from the book. This network is based off of Andrej Karpathy's post on RNNs and implementation in Torch; also, some information here at r2rt and from Sherjil Ozair on GitHub.
Below is the general architecture of the character-wise RNN.