Adaption of kernel optimizer

sQUlearn / squlearn

scikit-learn interface for quantum algorithms

https://squlearn.github.io

Apache License 2.0

58 stars 18 forks source link

Adaption of kernel optimizer #279

Closed DennisKleinhans closed 2 months ago

DennisKleinhans commented 4 months ago

The KernelOptimizer extends the KernelMatrixBase so that it can be used as a kernel in the high-level methods.
The KernelOptimizer now accepts a KernelMatrix, a loss function and an optimizer.
The KernelOptimizer automatically optimizes the provided KernelMatrix when the fit method is called by checking the is_trainable property.

closes #274

David-Kreplin commented 3 months ago

I did some adjustments, maybe you can check them @MoritzWillmann

David-Kreplin commented 3 months ago

I’ll take it from here. Most of the work looks great, but there are a few specifics that are a bit tedious to explain and fix, such as regularization.