Multiple models trained simultaneously on 1 solution

[efthimia145]

Hello,

I am using the PyGAD module to run a genetic algorithm that aims to find the optimal combination of image augmentations to use during the training phase of a model, in order to create a robust classification model. So I am not trying to optimize weights or parameters of the model- but only the combination of augmentations.

Each solution in the genetic algorithm is an array with elements that represent a specific augmentation and can have the value of 0 or 1 indicating whether an augmentation is applied or not. 

Therefore for each solution, I should have 1 model trained, which I then evaluate and the evaluation accuracy is my fitness function return.

❗As the genetic algorithm goes on and the solutions increment, the models trained on the specific solution are also incremented. e.g.

Solution 1 -> 1 Model is trained based on the solution
Solution 2 -> 2 Models are trained simultaneously on the same solution
Solution 3 -> 3 Models are trained simultaneously on the same solution
and so on..

It is like every time an additional process is starting and training a model on the same solution.

I know it might be a bit confusing, but I'd really appreciate any thoughts. I attach a python file with my code in case it's helfpul. sid_ga.zip

Thank you in advance!