I just stumbled across this which is really cool: https://cfwebprod.sandia.gov/cfdocs/CompResearch/docs/Xyce-PyMi-Guide.pdf Python based models inside Xyce. This totally removes any disappointment about lack of VerilogAMS support because this is so much better. Great job Xyce team

Yes, that is a new feature that we’re fairly proud of. It grew out of a project where people were developing data-based models that used Python-based machine learning libraries like TensorFlow and PyTorch. The researchers on that project would be working in Python anyway, and develop models based on these ideas, and then would ask, “can we run these in Xyce?” So, setting up this interface was the solution. It is pretty cool.

@User This is interesting ^^