REINFORCE

This repo. contains a reimplementation of the REINFORCE algorithm (vanilla policy gradient)

Clone the repository

git clone https://github.com/FaisalAhmed0/REINFORCE

You can setup a new environment and install requirements.txt

conda create -n vpg_env 
pip3 install -r requirements.txt 

You can set up the configuration of the experiment in congif.py

Activate the new environment and run train.py and pass the environment name as a cmd argument

conda activate vpg_env
python train.py --env "CartPole-v0"

to track the training dynamics run tensorbaord on the same directory

tensorboard --logir ./runs

after the training finsh you can run a video on the videos folder