Skip to content

LampV/vvlab

Repository files navigation

VVLAB

VVLAB is a reinforcement learning platform based on Pytorch and OpenAI Gym.The supported interface algorithms currently include:

  • Sarsa

  • Qlearning

  • Deep Q-Network (DQN)

  • Deep Deterministic Policy Gardient (DDPG)

Installation

Notice: The platform uses pytorch and numpy. It is recommended to use conda to create a new environment and install it.

VVLAB is currently hosted on PyPI. It requires Python >= 3.6.

You can simply install VVLAB from PyPI with the following command:

pip install vvlab

You can also install with the newest version through GitHub:

pip install git+https://github.com/LampV/Reinforcement-Learning.git@master

Or install it after downloading it locally:

git clone https://github.com/LampV/Reinforcement-Learning

Enter folder and install it with pip:

cd Reinforcement-Learning
pip install .

After installation, run examples :

python examples/ddpg.py

If no error occurs, you have successfully installed VVLAB.

Documentation

Todo

Quick Start

You can create your own reinforcement learning agent through the base class provided in vvlab.agents.The general method is as follows:

# import base class
from vvlab.agents import xxxBase
# inherit and complete necessary methods
class myxxx(xxxBase):  
    def _build_net(self):
        pass

learn more about the usage, by codes examples and annotated documentations under examples/ .

Simple neural networks built on pytorch are also provided in vvlab.models. You can take it as a simple implement of DRL neural network.

from vvlab.models import SimpleDQNNet

Third, if you want to call the attached envs, you should run __init__.py to register to gym. After that you can use standard gym methods to create it.

import vvlab  
env = gym.make('Maze-v0')

Contributing

VVALB is still under development. More algorithms and features are going to be added and we always welcome contributions to help make VVLAB better. If you would like to contribute, please check out this link.

About

No description, website, or topics provided.

Resources

License

Security policy

Stars

Watchers

Forks

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages