Simple PyTorch deep reinforcement algorithm implementation tested by OpenAI Gym. Easy for beginner š„³.
To install Gym, you can follow this instruction.
Since some environments like LunarLanderContinuous-V2
need fully installationļ¼you may need to additionally do this:
pip install 'gym[box2d]'
- By run
LunarLanderContinuous-V2_train.py
you can train a lunar landing robot by using SAC algorithm. - By run
LunarLanderContinuous-V2_test.py
you can see the performance of my trained robot after about 180 episodes.
This implementation is well arranged and easy to understand. Enjoy your learning of Deep Reinforcement Algorithm. For more implementations you can see spinning up which I followed.