Arraymancer kind of works with NLP as well. Here is an example of training a neural network to generate Shakespear-like text (or Jane Austin or if you feed it Nim code it will generate Nim-like text):
[https://github.com/mratsim/Arraymancer/blob/master/examples/ex06_shakespeare_generator.nim#L416](https://github.com/mratsim/Arraymancer/blob/master/examples/ex06_shakespeare_generator.nim#L416) The main issue right now is that the faster your CPU is, the slower the training ¯\\_(ツ)_/¯
