PyTorch implementation of a character-level RNN which learns to generate text after being fed a large training corpus.
The model is based Andrej Karpathy's blog post on the effectiveness of recursive neural networks.
The code is available under the permissive MIT license.