Swinging up a pendulmn by PPO(Proximal Policy Optimization)
Please refer to this Qiita entry.
python run_pendulumn.py train
python run_pendulumn.py replay
If you need movie (animation gif) file, set RECORD_MOVIE=True in the script run_pendulmn.py to output serial numbered png file and exec:
cd movie
./conv.sh
python plot_log.py
MIT