Reinforcement-Learning-with-PyTorch

Inspired by MorvanZhou, we implement a set of classic reinforcement learning algorithms (Q-Learning, SARSA, DQN, DDPG, A3C, PPO, etc) with PyTorch. To adapt to legacy runtime environments, it is recommended to create a virtual environment based on Python 3.6 using conda.

Requirements

To run the code properly, need to install the following certain packages:

• gym=0.16
• matplotlib=3.3.4
• numpy=1.19.5
• pandas=1.1.5
• torch=1.10.2

Getting Started

To train an AI model with a certain algorithm, e.g., DQN, run the following script from the contents/5_Deep_Q_Network directory:

python run_this.py