At present, there are separate implementations of the numpy-based and theano-based astrophysical and detector models which requires twice as much work if someone wants to add a new model or wants to change how a current model works. It will be a bit tricky to write things in a way that is fast for numpy-based models but also works for theano-based models (for-loops aren't slow for the latter since it gets compiled, but for-loops can be quite slow for the former), but I think for most cases it should be quite doable. Once we've worked out how to do it well for one of them, it should be very easy to extrapolate to the other functions
Code of Conduct
[X] I agree to follow this project's Code of Conduct
Instrument
Light curve fitting (Stages 4-6)
What is your suggestion?
At present, there are separate implementations of the numpy-based and theano-based astrophysical and detector models which requires twice as much work if someone wants to add a new model or wants to change how a current model works. It will be a bit tricky to write things in a way that is fast for numpy-based models but also works for theano-based models (for-loops aren't slow for the latter since it gets compiled, but for-loops can be quite slow for the former), but I think for most cases it should be quite doable. Once we've worked out how to do it well for one of them, it should be very easy to extrapolate to the other functions
Code of Conduct