A port of the example in micrograd
using Haskell
and ad
. Described in detail in my blog post.
The neural-network implementation is in MLP.hs
, and the demo in
micrograd-ad.hs
. The rest is code to
generate the graphics.
Install nix. Then, issue make
in the root of the repository.
The data will be generated and plotted, with the output in the output
directory.