文件列表:

build.sh (286, 2018-04-18)
CMakeLists.txt (1594, 2018-04-18)
data\figure\action_space.png (4157, 2018-04-18)
data\figure\logo.png (9159, 2018-04-18)
data\figure\observation_space.png (38333, 2018-04-18)
data\font_8x8\basic.txt (6016, 2018-04-18)
data\pursuit_model\predator\tfdqn_0.data-00000-of-00001 (5070352, 2018-04-18)
data\pursuit_model\predator\tfdqn_0.index (1701, 2018-04-18)
data\pursuit_model\predator\tfdqn_0.meta (8838928, 2018-04-18)
data\pursuit_model\prey\tfdqn_0.data-00000-of-00001 (3579408, 2018-04-18)
data\pursuit_model\prey\tfdqn_0.index (1688, 2018-04-18)
data\pursuit_model\prey\tfdqn_0.meta (8510319, 2018-04-18)
doc\get_started.md (5212, 2018-04-18)
examples\api_demo.py (2019, 2018-04-18)
examples\show_arrange.py (699, 2018-04-18)
examples\show_battle_game.py (332, 2018-04-18)
examples\train_against.py (9190, 2018-04-18)
examples\train_arrange.py (17415, 2018-04-18)
examples\train_battle.py (7958, 2018-04-18)
examples\train_battle_game.py (9203, 2018-04-18)
examples\train_gather.py (12437, 2018-04-18)
examples\train_multi.py (10600, 2018-04-18)
examples\train_pursuit.py (5713, 2018-04-18)
examples\train_single.py (7284, 2018-04-18)
examples\train_tiger.py (6540, 2018-04-18)
examples\train_trans.py (8723, 2018-04-18)
LICENSE (1077, 2018-04-18)
python\magent\builtin\common.py (1165, 2018-04-18)
python\magent\builtin\config\battle.py (943, 2018-04-18)
python\magent\builtin\config\double_attack.py (1213, 2018-04-18)
python\magent\builtin\config\forest.py (963, 2018-04-18)
python\magent\builtin\config\pursuit.py (904, 2018-04-18)
python\magent\builtin\config\__init__.py (0, 2018-04-18)
python\magent\builtin\mx_model\a2c.py (10630, 2018-04-18)
python\magent\builtin\mx_model\base.py (1779, 2018-04-18)
python\magent\builtin\mx_model\dqn.py (15724, 2018-04-18)
python\magent\builtin\mx_model\__init__.py (68, 2018-04-18)
python\magent\builtin\rule_model\random.py (495, 2018-04-18)
... ...

  MAgent is a research platform for many-agent reinforcement learning. Unlike previous research platforms that focus on reinforcement learning research with a single agent or only few agents, MAgent aims at supporting reinforcement learning research that scales up from hundreds to millions of agents. - AAAI 2018 demo paper: [MAgent: A Many-Agent Reinforcement Learning Platform for Artificial Collective Intelligence](https://arxiv.org/abs/1712.00600) - Watch [our demo video](https://www.youtube.com/watch?v=HCSm0kVolqI) for some interesting show cases. - Here are two immediate demo for the battle case. ## Requirement MAgent supports Linux and OS X running Python 2.7 or python 3. We make no assumptions about the structure of your agents. You can write rule-based algorithms or use deep learning frameworks. ## Install on Linux ```bash git clone git@github.com:geek-ai/MAgent.git cd MAgent sudo apt-get install cmake libboost-system-dev libjsoncpp-dev libwebsocketpp-dev bash build.sh export PYTHONPATH=$(pwd)/python:$PYTHONPATH ``` ## Install on OSX ```bash git clone git@github.com:geek-ai/MAgent.git cd MAgent brew install cmake llvm boost brew install jsoncpp argp-standalone brew tap david-icracked/homebrew-websocketpp brew install --HEAD david-icracked/websocketpp/websocketpp bash build.sh export PYTHONPATH=$(pwd)/python:$PYTHONPATH ``` ## Docs [Get started](/doc/get_started.md) ## Examples The training time of following tasks is about 1 day on a GTX1080-Ti card. If out-of-memory errors occur, you can tune infer_batch_size smaller in models. **Note** : You should run following examples in the root directory of this repo. Do not cd to `examples/`. ### Train Three examples shown in the above video. Video files will be saved every 10 rounds. You can use render to watch them. * **pursuit** ``` python examples/train_pursuit.py --train ``` * **gathering** ``` python examples/train_gather.py --train ``` * **battle** ``` python examples/train_battle.py --train ``` ### Play An interactive game to play with battle agents. You will act as a general and dispatch your soldiers. * **battle game** ``` python examples/show_battle_game.py ``` ## Baseline Algorithms The baseline algorithms parameter-sharing DQN, DRQN, a2c are implemented in Tensorflow and MXNet. DQN performs best in our large number sharing and gridworld settings. ## Acknowledgement Many thanks to [Tianqi Chen](https://tqchen.github.io/) for the helpful suggestions.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：