Hacker News Reader

Also, correct me if I'm wrong, but the state of the art seems to be happening in Tensorflow/Keras; so committing to a different platform could mean you are systematically lagging behind in this new field.

Then don't use Haskell! I'm having a hard time understanding if your question is actually genuine or if you're just dropping by the Haskell thread to explain why you're not into using Haskell.

We have to remember that machine learning is much more than just ANN, or, for that sake, matrix manipulation:

http://probabilistic-programming.org/wiki/Home

You can for example prevent "runtime" bugs due to combining layers/vectors/matrixes of the wrong dimensions.

regarding compiling

>Grenade layers are normal haskell data types which are an instance of Layer, so it's easy to build one's own downstream code. We do however provide a decent set of layers, including convolution, deconvolution, pooling, pad, crop, logit, relu, elu, tanh, and fully connected.

it's called a README for a reason.