I haven't looked at the code, but glancing at the results leaves me thinking it might need more work.
The output seems to me around the level a Markov chain might produce. Karpathy's RNN code produces much, much better results[1].
I wonder if manually extracting features and training the RNN on that is a mistake? RNN's tend to work well on text because they encode understanding of the parse tree themselves.
Comments
I haven't looked at the code, but glancing at the results leaves me thinking it might need more work.
The output seems to me around the level a Markov chain might produce. Karpathy's RNN code produces much, much better results[1].
I wonder if manually extracting features and training the RNN on that is a mistake? RNN's tend to work well on text because they encode understanding of the parse tree themselves.
[1] https://github.com/karpathy/char-rnn