DINS Reinforced Learning Workshop project

This is testing RL environment for DINS RL Workshop. Repository contains basic scaffold and setup guide.

Setup

Environment can be setup on Linux, MacOS or Windows (using linux subsystem) machines. Guide assumes Ubuntu 16.04-LTS is used.

For Windows, XServer should be installed (like Xming)

Python 3 should be installed on machine as pre-requisite

Step 1

Install Python Virtual Env.

sudo apt-get install virtualenv
pip install virtualenv
pip install virtualenvwrapper
export WORKON_HOME=~/Envs
source /usr/local/bin/virtualenvwrapper.sh
mkdir -p $WORKON_HOME
# Create virtual env
mkvirtualenv dins-workshop

Step 2

Install dependencies:

# absl
pip install absl-py
# Numpy
pip install numpy
# OpenCV Python
pip install opencv-python
# OpenAI gym
git clone https://github.com/openai/gym.git
cd gym
pip install -e .

Step 3

Install OpenAI baselines

sudo apt-get install libcr-dev libopenmpi-dev mpich mpich-doc zlib1g-dev cmake python-opencv 
git clone https://github.com/openai/baselines.git
cd baselines
pip install -e .

Step 4

Install fceux and super-mario-gym environment

sudo apt-get install fceux
git clone https://github.com/yoks/gym-super-mario.git
pip install -e .

Training

python train.py

Running

python run.py