This implementation is inspired by OpenAI spinning up.
The implementation of DDPG only contains one
script: DDPG/main.py.
To train a DDPG agent, write ddpg(env_name="HalfCheetah-v2")
in main function of DDPG/main.py.
To play with pre-trained model, write
play(model_path="model/model_60.h5") in the main
function of DDPG/main.py
The implementation of TD3 only contains one
script: TD3/main.py.
To train a TD3 agent, write td3(env_name="HalfCheetah-v2")
in main function of TD3/main.py.
To play with pre-trained model, write
play(model_path="model/model_30.h5") in the main
function of TD3/main.py.