Has anybody tried to use grid search for tuning the PennyLane model hyper-parameters, especially when using Torch interface. There is a torch wrapper Skorch around sci-learn to automate the tuning process. Any suggestions? (it worries me that going down this path we get too many dependencies)
Hi @Jacob_Cybulski , welcome to the Forum!
Some members of our team have done experiments with hyperparameter optimization, but I’m not aware of what tools they’ve used in the past.
Last year our QML team also published some very interesting work on benchmarking quantum machine learning models. If you want to learn more about their work you can start by reading the blog post, and then explore the paper and the GitHub repository.
I hope this helps!
Thank you Catalina, it is a good start!
1 Like