Hands-On Neuroevolution with Python

PyTorch NEAT usage example

The following is an example of using the PyTorch NEAT library to implement a cartpole balancing controller. This is only being presented as an overview. Later in this book, we will dive deeper into the pole balancing problem in more detail. Let's get started:

1. Load the NEAT settings and seed genome configuration:

config = neat.Config(neat.DefaultGenome, neat.DefaultReproduction,       neat.DefaultSpeciesSet, neat.DefaultStagnation, config_file)

Here, the config_file file stores the NEAT algorithm settings, along with the default genome configuration.

2. Create a population of organisms from the configuration data:

pop = neat.Population(config)

3. Prepare the multi-environment genome evaluator based on PyTorch and OpenAI ...