README.md

work in progress repo

BasicRL

A basic implementation of the standard reinforcement learning algorithms in Pytorch, designed to work with gym-like environments. All others algorithms in Tensorflow, basicRl.py and Myplotter.py have been written by Davide Corsi

Available Algorithms in Pytorch

• DQN
• REINFORCE
• PPO
• DDPG

Available Algorithms in TensorFlow (made by D.Corsi)

• REINFORCE
• Actor-Critic
• Advantage Actor Critic (A2C)
• Proximal Policy Optimization (PPO)
• Montecarlo Proximal Policy Optimization (mcPPO)
• Double Deep Q-Learning (DDQN)
• Deep Deterministic Policy Gradient (DDPG)
• Twin Delayed DDPG (TD3)

Prerequisites

• Python

Libraries and Frameworks:

pip install gym=0.26
pip install tensorflow
pip install pytorch
pip install matplotlib
pip install numpy 

Run the Algorithms

Download the repo:

  git clone https://github.com/Enange/BasicRLMaster.git

Run the main script

  python example.py

Built With

• Pytorch
• Tensorflow
• OpenAI Gym
• Numpy
• Matplot Lib

Author

• Davide Corsi -davide.corsi@univr.it
• Enrico Angelico
• Riccardo Portaro