Hyperparameter Grid Search Pytorch

Hi everyone,

I was wondering if there is a simple way of performing grid search for hyper-parameters in pytorch? For example, assuming I have 3 possible values for parameter a, 3 for param b and 4 for param c, I have a total of 3 * 3 * 4 = 36 different combinations of hyper-parameters. Is there any way to do this efficiently? Or any external library which is integrated with PyTorch to perform this search.

My motivation: Hyperparameters can drastically change the performance of a model. This seems like a very scary prospect since you can never know if your model is actually performing well or it is failing due to some hyper parameter combinations…
Please let me know if anyone has a solution for easy hyperparameter search.
Thank you!

Check out this thread:

Thank you, I will take a look at it.