The implementation of Multistep-DDPG and Mixed-Multistep-DDPG proposed in The Effect of Multi-step Methods on Overestimation in Deep Reinforcement Learning.
Install Multistep-DDPG and dependencies.
git clone https://github.com/LinghengMeng/Multistep-DDPG.git
cd Multistep-DDPG
pip install -e .
To use tasks from PyBulletGym, please follow the instructions in https://github.com/benelot/pybullet-gym to install Pybullet-Gym.
python .\mddpg\mddpg_main.py
The code for other baselines:
- DDPG, SAC, TD3 can be found in Spinningup
- MVE, STEVE can be found in Stochastic-Ensemble-Value-Expansion.