-
Notifications
You must be signed in to change notification settings - Fork 42
/
README.md.orig
44 lines (27 loc) · 3.16 KB
/
README.md.orig
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
[](http://rllab.readthedocs.org/en/latest/)
[](https://circleci.com/gh/rllab/rllab)
[](https://github.com/rllab/rllab/blob/master/LICENSE)
[](https://gitter.im/rllab/rllab?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
# rllab
rllab is a framework for developing and evaluating reinforcement learning algorithms. It includes a wide range of continuous control tasks plus implementations of the following algorithms:
- [REINFORCE](https://github.com/rllab/rllab/blob/master/rllab/algos/vpg.py)
- [Truncated Natural Policy Gradient](https://github.com/rllab/rllab/blob/master/rllab/algos/tnpg.py)
- [Reward-Weighted Regression](https://github.com/rllab/rllab/blob/master/rllab/algos/erwr.py)
- [Relative Entropy Policy Search](https://github.com/rllab/rllab/blob/master/rllab/algos/reps.py)
- [Trust Region Policy Optimization](https://github.com/rllab/rllab/blob/master/rllab/algos/trpo.py)
- [Cross Entropy Method](https://github.com/rllab/rllab/blob/master/rllab/algos/cem.py)
- [Covariance Matrix Adaption Evolution Strategy](https://github.com/rllab/rllab/blob/master/rllab/algos/cma_es.py)
- [Deep Deterministic Policy Gradient](https://github.com/rllab/rllab/blob/master/rllab/algos/ddpg.py)
rllab is fully compatible with [OpenAI Gym](https://gym.openai.com/). See [here](http://rllab.readthedocs.io/en/latest/user/gym_integration.html) for instructions and examples.
rllab only officially supports Python 3.5+. For an older snapshot of rllab sitting on Python 2, please use the [py2 branch](https://github.com/rllab/rllab/tree/py2).
rllab comes with support for running reinforcement learning experiments on an EC2 cluster, and tools for visualizing the results. See the [documentation](https://rllab.readthedocs.io/en/latest/user/cluster.html) for details.
The main modules use [Theano](http://deeplearning.net/software/theano/) as the underlying framework, and we have support for TensorFlow under [sandbox/rocky/tf](https://github.com/openai/rllab/tree/master/sandbox/rocky/tf).
# Documentation
Documentation is available online: [https://rllab.readthedocs.org/en/latest/](https://rllab.readthedocs.org/en/latest/).
# Citing rllab
If you use rllab for academic research, you are highly encouraged to cite the following paper:
- Yan Duan, Xi Chen, Rein Houthooft, John Schulman, Pieter Abbeel. "[Benchmarking Deep Reinforcement Learning for Continuous Control](http://arxiv.org/abs/1604.06778)". _Proceedings of the 33rd International Conference on Machine Learning (ICML), 2016._
# Credits
rllab was originally developed by Rocky Duan (UC Berkeley / OpenAI), Peter Chen (UC Berkeley), Rein Houthooft (UC Berkeley / OpenAI), John Schulman (UC Berkeley / OpenAI), and Pieter Abbeel (UC Berkeley / OpenAI). The library is continued to be jointly developed by people at OpenAI and UC Berkeley.
# Slides
Slides presented at ICML 2016: https://www.dropbox.com/s/rqtpp1jv2jtzxeg/ICML2016_benchmarking_slides.pdf?dl=0