gym-Atari-SpaceInvaders-V0

The instruction is only for Ubuntu.

Neon

Install prerequisites:

sudo apt-get install libhdf5-dev libyaml-dev libopencv-dev pkg-config
sudo apt-get install python python-dev python-pip python-virtualenv

Check out and compile the code:

git clone https://github.com/NervanaSystems/neon.git
cd neon
make

Set Python PATH to NEON_HOME. NEON_HOME is the path where neon is installed,

export PYTHONPATH = NEON_HOME

OpenAI Gym

Install OpenAI Gym:

pip install gym
pip install gym[atari]

Simple DQN

Prerequisities:

pip install numpy argparse logging

To install openCV

sudo apt-get install python-opencv
ln -s /usr/lib/python2.7/dist-packages/cv2.so NEON_HOME/.venv/lib/python2.7/site-packages/
In case it doen't work try following,

echo "deb http://archive.ubuntu.com/ubuntu/ trusty main restricted universe multiverse" >> /etc/apt/sources.list
apt-get update

apt-get -y install libopencv-dev build-essential cmake git libgtk2.0-dev pkg-config python-dev python-numpy libdc1394-22 libdc1394-22-dev libjpeg-dev libpng12-dev libtiff4-dev libjasper-dev libavcodec-dev libavformat-dev libswscale-dev libxine-dev libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev libv4l-dev libtbb-dev libqt4-dev libfaac-dev libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev x264 v4l-utils unzip

cd /tmp
git clone https://github.com/opencv/opencv.git
cd opencv
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE\
-D CMAKE_INSTALL_PREFIX=/usr/local\
-D WITH_TBB=ON \ 
-D WITH_V4L=ON \
-D WITH_QT=ON \
-D WITH_OPENGL=ON ..

make -j $(nproc)
make install

/bin/bash -c 'echo "/usr/local/lib" > /etc/ld.so.conf.d/opencv.conf'
ldconfig

ln /dev/null /dev/raw1394

cd /
rm -rf /tmp/opencv

Then just check out the code:

git clone https://github.com/tambetm/simple_dqn
cd simple_dqn

Optional

For plotting install matplotlib:

pip install matplotlib

For producing game videos install avconv:

sudo apt-get install libav-tools

To save video files

Replace simple_dqn/agent.py in the current code with agent.py in the current repository. Update (the location where the video files will be saved) line 138

Running the code

For training, testing and saving the results in a csv file

python main.py game SpaceInvaders-V0 --environment "gym" --backend 'cpu' --random_steps 500 --train_steps 1000 --test_steps 500 --csv_file 'SpaceInvaders-V0.csv'

Note :

Some trivial issues may arise due to different python versions. This is compatible for Python2. If you want to run it using python3, you have to change the xrange to range and sys.maxint to sys.maxsize

Some useful links :

1. https://github.com/tambetm/simple_dqn/tree/646154a147872e653873244fd43bc04490d99457
2. https://gym.openai.com/evaluations/eval_Eduozx4HRyqgTCVk9ltw
3. https://github.com/ppwwyyxx/tensorpack/tree/master/examples/OpenAIGym