Policy Gradient

Implementation of policy gradients in Pytorch.

Getting Started

Install the following prerequisites on your system

• pytorch
• torchvision
• opencv
• gym
• gym[atari]

To execute a policy gradient, run the main.py file.

python main.py

All of the policy gradient training and optimizer parameters are at the top of main.py so feel free to modify these to suit your configuration.

There are some parameter configurations on the command line. More will be added!

Todos

• More command line configurations
• Train for a few days and post results

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Berkeley Deep RL course for DeepMind Atari wrappers and program structure (https://github.com/berkeleydeeprlcourse/homework)
• Pytorch examples (https://github.com/pytorch/examples/blob/master/reinforcement_learning/reinforce.py)