We are trying to give a model to collaborators but would like to protect the IP. What are some ways to encrypt/hide/compile the definition when sharing a trained model?
1 Answer
One way to protect the IP of a trained model is to put the model behind an API. Give only limited API end-point access.
-
$\begingroup$ Yes, something like docker. I did think of that one. Is there no way to "compile" a model to something like a binary? $\endgroup$– illanFeb 8 at 22:18