lexiconium/rl-gym-pytorch

Implement RL algorithms in PyTorch and test on Gym environments.

RL-Gym-PyTorch

The purpose of this repository is to implement Reinforcement Learning algorithms in PyTorch and test them on a variety of OpenAI Gym environments.

All implementations are specific to each environment with minimum generalization so that the entire structure of the algorithm can be seen as clearly as possible.

Environments

Each directory contains:

• Overview and Specifications of the environment
• Algorithms*

*: Includes pseudocode, code and result

Availability

• CartPole-v1
  • DQN
  • PPO
• Pendulum-v0
  • DDPG
• LunarLander-v2
  • PPO (Continuous)
  • PPO (Discrete)
  • TD3 (Continuous)
• MountainCarContinuous-v0
  • DDPG
• Walker2d
  • TD3

Dependencies

Python 3.8.12
Gym 0.19.0
mujoco-py 2.1.2.14
PyTorch 1.10.1
NumPy 1.21.5