This project is still a work in progress. More algorithms and detailed documentation coming soon :)
Currently supported agents-
- Random agent
- REINFORCE (Policy Gradients)
- DQN
- DQN with baseline
- Actor-Critic
See examples for details on how to use the library.
$ pip install -U RLkit
To get the project's source code, clone the github repository:
$ git clone https://github.com/shubhamjha97/rlkit.git
$ cd RLkit
Install VirtualEnv using the following (optional):
$ [sudo] pip install virtualenv
Create and activate your virtual environment (optional):
$ virtualenv venv
$ source venv/bin/activate
Install all the required packages:
$ pip install -r requirements.txt
Install the package by running the following command from the root directory of the repository:
$ python setup.py install
- Added DQN and DQN with baseline agents
- Added ActorCritic agent
- Added support for various activation functions
- Support for OpenAI Gym environments, Vizdoom and custom environments
- Generic constructs such as Environments, Agent and Trainers
- Better Logging and tracking of metrics
- Support for CometML
- Support for the following algorithms
- Duelling DQN
- Support for logging and plotting
- Support for adding seeds
The package has been tested with python 3.5.2
-
Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., & Riedmiller, M. (2013). Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602. [paper]
-
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., ... & Petersen, S. (2015). Human-level control through deep reinforcement learning. Nature, 518(7540), 529. [paper]
-
Williams, R. J. (1992). Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning, 8(3-4), 229-256. [paper]
-
Konda, V. R., & Tsitsiklis, J. N. (2000). Actor-critic algorithms. In Advances in neural information processing systems (pp. 1008-1014). [paper]