=============

• python 2.7

Install Unity (experimental-build version):

wget http://download.unity3d.com/download_unity/unity-editor-installer-5.1.0f3+2015082501.sh
sudo sh unity-editor-installer-5.1.0f3+2015082501.sh

# run Unity
./unity-editor-5.1.0f3/Editor/Unity

# if background is pink, install:
sudo apt-get install lib32stdc++6 -y

install python modules:

pip install chainer
pip install ws4py
pip install cherrypy
pip install msgpack-python

Install Unity. (if you are going to use Ubuntu for GPU pwer, I reccomend install Unity 5.1.0 linux experimental-build version)

install python modules:

pip install chainer
pip install ws4py
pip install cherrypy
pip install msgpack-python

超人工生命ハッカソンに向けてWindows10マシンにpython環境を整えた

download data:

./fetch.sh

Next, run python module as a server.

cd python-agent
python server.py

Open unity-sample-environment in Unity and load Scenes/sample.

Press Start Buttn. This will take a few minuts for loading caffe model.

You can watch reward history:

cd python-agent
python plot_reward_log.py

This graph is a "sample" scene result. It takes about 6 hours on GPU Machine.

Sample scene result movie

SampleLikesAndDislikes scene result movie

• Client: Unity
• Server: python module
• Communication: Socket (WebSocket over TCP) using MessagePack

Set "Time Scale" to 100. This will make simulation more faster, but GUI response will be very slow.

• Mnih, V. et al. Human-level control through deep reinforcement learning. Nature 518, 529–533 (2015)

• http://www.nature.com/nature/journal/v518/n7540/abs/nature14236.html

• DQN-chainer

• A. Krizhevsky, I. Sutskever, and G. Hinton. ImageNet classification with deep convolutional neural networks. In NIPS, 2012.

• Caffe Model Zoo

• MessagePack for Unity

• Copyright (C) 2011-2012 Kazuki Oikawa, Kazunari Kida

• Apache License, Version 2.0

• Assets/Packages/msgpack-unity

• websocket-sharp

• Copyright (c) 2010-2016 sta.blockhead

• The MIT License (MIT)

• Assets/Packages/websocket-sharp

• Apache License, Version 2.0
• This framework was made by DWANGO ARTIFICIAL INTELLIGENCE LABORATORY and lab member Masayoshi Nakamura

If you created intelligent agents, please let me know about it to "masayoshi_nakamura@dwango.co.jp". We will make showcase varied intelligent agents.