ACER

KO | EN

gitlite — search

#javascript #python #typescript #go #golang #machine-learning #macos #nodejs #kubernetes #awesome-list #security #openai #ios #open-source

(★ 257)

Actor-critic with experience replay

파일 탐색기

최종 버전 다운로드 (.zip)

- acer.png
.gitignore
CONTRIBUTING.md
environment.yml
LICENSE.md
main.py
memory.py
model.py
optim.py
README.md
test.py
train.py
utils.py

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

ACER
====
[![MIT License](https://img.shields.io/badge/license-MIT-blue.svg)](https://raw.githubusercontent.com/Kaixhin/ACER/master/LICENSE.md)

Actor-critic with experience replay (ACER) [[1]](#references). Uses batch off-policy updates to improve stability. Trust region updates can be enabled with `--trust-region`. Currently uses full trust region instead of "efficient" trust region (see [issue #1](https://github.com/Kaixhin/ACER/issues/1)).

Run with `python main.py <options>`. To run asynchronous advantage actor-critic (A3C) [[2]](#references) (but with a Q-value head), use the `--on-policy` option.

Requirements
------------

- [OpenAI Gym](https://gym.openai.com/)
- [Plotly](https://plot.ly/python/)
- [PyTorch](http://pytorch.org/)

To install all dependencies with Anaconda run `conda env create -f environment.yml` and use `source activate acer` to activate the environment.

Results
-------

![ACER](https://raw.githubusercontent.com/Kaixhin/ACER/master/figures/acer.png)

Acknowledgements
----------------

- [@ikostrikov](https://github.com/ikostrikov) for [pytorch-a3c](https://github.com/ikostrikov/pytorch-a3c)
- [@apaszke](https://github.com/apaszke) for [Reinforcement Learning (DQN) tutorial](http://pytorch.org/tutorials/intermediate/reinforcement_q_learning.html)
- [@pfnet](https://github.com/pfnet) for [ChainerRL](https://github.com/pfnet/chainerrl)

References
----------

[1] [Sample Efficient Actor-Critic with Experience Replay](https://arxiv.org/abs/1611.01224)  
[2] [Asynchronous Methods for Deep Reinforcement Learning](https://arxiv.org/abs/1602.01783)

ACER

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

ACER

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README