Packaging trained model

Hi I trained a model and deployed the model as a Django app. I am trying to make the Django app as a package using setuptools which works however I can’t understand how to include pytorch as a requirement. If i put fastai it doesn’t work as I need to run the model on CPU which means I have to install torch separately but pip can’t find the cpu version of torch.

I can manually install the cpu version of torch with pip3 install torch==1.2.0+cpu torchvision==0.4.0+cpu -f but I can’t seem to understand how to do the same with setuptools. I have tried using dependency links but I just can’t seem to get anything working and I couldn’t find anything about this online. Has anyone packaged pytorch cpu as a requirement with setuptools and how did you do it?