add script for training qmix or mqmix

render result in qmix/rnn algorithm

update tiny, update render.py

update tiny for training mpe env

feature: enable to render result

fix module import

update env_core.py, adjust action space

fix custom env, enable to train in qmix algorithm