KaleabTessera / Policy-Gradient

Implementation of the following Policy Gradient Algorithms -Reinforce and Actor Critic.

Policy Gradient

Implementation of the following Policy Gradient Algorithms - Reinforce and Actor Critic. Some parts adapted from https://github.com/andrecianflone/rl_at_ammi .

Reinforce

Reinforce vs Reinforce with Baseline

Video of Gameplay - Reinforce Algorithm for the Lunar-Lander environment

Actor-Critic

Video of Gameplay - Actor-Critic implementation for Bipedal walker

How to run

Create a new environment

Example:

conda create -n policy_grad

Install Dependencies

pip install -r requirements.txt

To use gym.wrappers.Monitor to record the last episode

sudo apt-get install ffmpeg

Reinforce

Without baseline

python reinforce.py

With baseline

python reinforce_with_learned_baseline.py

Actor-Critic

Run Training from Scratch

python actor_critic.py

Use a trained agent

python actor_critic.py --load-model