How can I use DEAP for Generating Neural Network Architecture ?

[Aditya Lokesh]

I came across this framework for genetic programming and would like some insight on how can I be able to use this framework to generate Neural network architecture.Since the type of input that can be specified for the problem is list,string ,array and so on can I just use the population type as a graph if so how can I do that?.

My project entails finding a neural network architecture using genetic algorithm.

If I cannot use graph as Input type is there support to generate vector of non-integral values representing a graph and use the vector instead and generate a population?

[Martin Karas]

Hello,

NEAT implements it:

https://neat-python.readthedocs.io/en/latest/

have you found DEAP implementation of generating NN? 

Dátum: streda 19. februára 2020, čas: 12:20:38 UTC+1, odosielateľ: aditya...@gmail.com