naiveRL is a concise implementation of mainstream reinforcement learning algorithms. It strives to use a unified code style and each folder contains an easy-to-use implementation of a corresponding algorithm. This makes it convenient to use and modify for academic purposes. Currently, the implemented algorithms include A2C, A3C, PPO, D3QN, DDPG, and SAC. The code mainly refers to the implementation of minimalRL and PARL.

Each algorithm is implemented in a separate folder. To use an algorithm, simply navigate to the corresponding folder and run the main script. For example, to run PPO, navigate to the PPO folder and run: python run.py The default environment is CartPole-v0, but you can change the environment by modifying the env_name parameter in the config.py file.

This project is licensed under the MIT License