Generative adversarial network in 70 lines of Julia

Neat that you can differentiate through functions passed using keyword arguments. Also the unicode letters except for delta shows up as black bars in my android which is kind of annoying.

That’s compelling as a demonstration of Julia’s expressiveness.

My next thought is: what would it look like in Python?

With and without a deep learning library.

I suspect it would be much shorter with and slightly longer without but I am not certain.

Knet is pretty analogous to pytorch/torch in terms of terseness. I’d expect about the same length them and a quite larger script with tensorflow (based on my very limited experience with it)

Doesn't seem to really take advantage of Julia's compelling tools but still neat.