This repo is an implementation of NASNET, which proposed from Neural Architecture Search with Reinforcement Learning. I think automl with RL can be a general method to generate models. But I can not find a github repo successfully implement layer anchor. So I Recreated the wheel.
some codes referenced this repo.
CPU i7-7700 Python 3.6.8 Pytorch 1.1.0
NAS model(NAS cell + reinforce framework)
CNN model
child network manager
python pyt_train.py