Unofficial PyTorch implementation of Neural Additive Models (NAM) by Agarwal, et al. [abs, pdf]

You can access nam-pytorch via pip:

$ pip install nam-pytorch

import torch 
from nam_pytorch import NAM

nam = NAM(
    num_features=784,
    link_func="sigmoid"
)

images = torch.rand(32, 784)
pred = nam(images) # [32, 1]

As always, if there are any issues with / suggestions for the code, feel free to raise an issue or submit a PR.

MIT