In this repo I train an agent using the reinforce algorithm in which I use two traces to estimate the graident of the policy. The agent was trained and tested over the enviroment CartPole-v0 from OpenAI-GYM.
The RL algorithm is located under jupyter notebook REINFORCE.ipynb in which you can train the agent from scratch using the default params or testing it with diferent configs.
To use this code you need to install the following packages:
- gym
- torch
- numpy
- jupiyter
- matplotlib
GNU General Public License v3.0