Learning rate and convergence threshold for cudaq's Adam optimizer

[anandaroy7]

Hello, Is it possible for one to tune the learning rate and the convergence threshold for the Adam optimizer?

Thanks.

[bmhowe23]

Hi @anandaroy7 - are you using C++ or Python? If you're using C++, I think you can use the f_tol and step_size attributes. I am not sure if these are exposed to Python at the moment, but if they are missing, we could certainly add them.

[anandaroy7]

Hi, I am using the python version and it is not practical for me to switch to C++. It would be great if you could add the ability to control the convergence threshold and the learning rate in the python version. Thanks!