Does this implementation accept multiple hidden layers?
valeman opened this issue · 1 comments
valeman commented
I was wondering if this implementation allowed for multiple hidden layers that is key in getting good performance.
csinva commented
I just pushed support for this :)
You'll need to update the package (pip install imodelsx
), then can pass hidden_layer_sizes
list for multiple layers (like sklearn's MLP). See the second cell of the updated notebook for an example.