- Original paper: https://arxiv.org/abs/1707.06347
- Baselines blog post: https://blog.openai.com/openai-baselines-ppo/
mpirun -np 8 python -m baselines.ppo1.run_atari
runs the algorithm for 40M frames = 10M timesteps on an Atari game. See help (-h
) for more options.python -m baselines.ppo1.run_mujoco
runs the algorithm for 1M frames on a Mujoco environment.
ppo1
Folders and files
Name | Name | Last commit date | ||
---|---|---|---|---|
parent directory.. | ||||