The purpose of this miniproject was to examine the NN proposed by Martin Görner in its functionality as well as its performance on different kinds of input datasets, other than the already tested ones (Shakespeare and Python Code). We used three kind of other data: Books of Kafka, Whatsapp conversations and song lyrics. By tweaking Görners code only as little as possible, we were able to make the NN work on our data flawlessly.
The used code is based on https://github.com/martin-gorner/tensorflow-rnn-shakespeare