train.py

This train.py module acts as our top-level entry to training the Controller:

import loggingimport sysfrom .controller import Controllerif __name__ == '__main__':    # Configure the logger    logging.basicConfig(stream=sys.stdout,                        level=logging.DEBUG,                        format='%(asctime)s %(name)-12s %(levelname)-8s %(message)s')    controller = Controller()    controller.train_controller()

And there we have it; a neural network that generates other neural networks! Make sure your implementation has the following directory structure:

src|-- __init__.py|-- child_network.py|-- cifar10_processor.py|-- config.py|-- constants.py|-- controller.py`-- train.py

To execute training, simply run the following command:

$ python train.py

If all works well, you should be seeing ...

Get Python Reinforcement Learning Projects now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.