Skip to content

Light for lightweight and fast of DRL

License

Notifications You must be signed in to change notification settings

i-am-neet/lightDRL

 
 

Repository files navigation

Light DRL

Install

  1. Download this package from git clone [email protected]:lightDRL/lightDRL.git or download this zip

  2. Install python requirements by following command

pip install -r requirements.txt

# if fail, try to use sudo 
sudo pip install -r requirements.txt

Run Standalone

Test example which you want to play, for example, play gym

python examples/gym_basic/gym_basic.py examples/gym_basic/DDPG_CartPole-v0.yaml

Or you can

cd examples/gym_basic
python gym_basic.py DDPG_CartPole-v0.yaml

Run with Server

python server.py
python examples/gym_basic/gym_basic_conn.py examples/gym_basic/DDPG_CartPole-v0.yaml

Or you can

cd examples/gym_basic
python ../../server.py
python gym_basic_conn.py DDPG.yaml

Version Note

v0.13 (latest)

Support NNcomponent, you can modify network by config.yaml

v0.121

Update Q-learning to more general Play maze with Q-learning Add cfg['misc']['ep_max_step'], if done = True and self.ep_use_step >= self.ep_max_step -> train_done = False

python examples/maze/maze.py  examples/maze/Q-learning.yaml

v0.12

Support DQN! Now support DQN, DDPG, and Q-learning

Play maze with DQN

python examples/maze/maze.py examples/maze/DQN.yaml

v0.11

Support Q-learning!

python examples/gym_basic/gym_basic.py examples/gym_basic/Q-learning_Taxi-v2.yaml

v0.10

Change framework

Now, support standalone & server-client version !!

About

Light for lightweight and fast of DRL

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%