Hi，
I added my own network efficiency predictor and accuracy predictor, and then specialized the network through the evolutionary algorithm. It was normal at the beginning, but after a certain number of iterations, the network structure of the child generation pool basically changed into the same one, and the accuracy no longer increased. May I ask why? Is it trapped in a locally optimal solution?
Hi， I added my own network efficiency predictor and accuracy predictor, and then specialized the network through the evolutionary algorithm. It was normal at the beginning, but after a certain number of iterations, the network structure of the child generation pool basically changed into the same one, and the accuracy no longer increased. May I ask why? Is it trapped in a locally optimal solution?