search

MarkoMlakar / sklearn-neuro-evolution

BSD 3-Clause "New" or "Revised" License
8 stars 0 forks source link

Dense matrices are required as an X input, which differs from all other classifiers #2

Open qwertyuu opened 4 years ago

qwertyuu commented 4 years ago

First of all, thanks a whole lot for making this sklearn package to easily add NEAT to sklearn!

I wanted to use your package quickly to apply NEAT to a problem that I had tried many different classifiers with before. The whole pipeline was setup and worked great, but the neuro evolution is the only one requiring dense matrices for an input.

I am just putting it out there that the sparse/dense matrix detection could be added to the package, a warning could be added before the fitting to inform that the input matrix will be converted to a dense one under the hood and let your wrapper take care of the rest by itself.

MarkoMlakar commented 4 years ago

Hello @qwertyuu thank you for the response ! I will look into it as soon as I have time :)