Acme – A framework for distributed reinforcement learning

栏目: IT技术 · 发布时间: 4年前

内容简介:||

Acme – A framework for distributed reinforcement learning

Acme: A research framework for reinforcement learning

|| Documentation | Agents | Examples | Paper

Acme is a library of reinforcement learning (RL) agents and agent building blocks. Acme strives to expose simple, efficient, and readable agents, that serve both as reference implementations of popular algorithms and as strong baselines, while still providing enough flexibility to do novel research. The design of Acme also attempts to provide multiple points of entry to the RL problem at differing levels of complexity.

Overview

At the highest level Acme exposes a number of agents which can be used simply as follows:

import acme

# Create an environment and an actor.
environment = ...
actor = ...

# Run the environment loop.
loop = acme.EnvironmentLoop(environment, actor)
loop.run()

Acme also tries to maintain this level of simplicity while either diving deeper into the agent algorithms or by using them in more complicated settings. An overview of Acme along with more detailed descriptions of its underlying components can be found by referring to the documentation .

For a quick start, take a look at the more detailed working code examples found in the examples subdirectory, which also includes a tutorial notebook to get you started. And finally, for more information on the various agent implementations available take a look at the agents subdirectory along with the README.md associated with each agent.

Installation

We support Python 3.6 and 3.7.

To install acme core:

# Install Acme core dependencies.
pip install dm-acme

# Install Reverb, our replay backend.
pip install dm-acme[reverb]

To install dependencies for our JAX/TensorFlow-based agents:

pip install dm-acme[tf]
# and/or
pip install dm-acme[jax]

Finally, to install environments (gym, dm_control, bsuite):

pip install dm-acme[envs]

Citing Acme

If you use Acme in your work, please cite the accompanying technical report :

@article{hoffman2020acme,
    title={Acme: A Research Framework for Distributed Reinforcement Learning},
    author={Matt Hoffman and
            Bobak Shahriari and
            John Aslanides and
            Gabriel Barth-Maron and
            Feryal Behbahani and
            Tamara Norman and
            Abbas Abdolmaleki and
            Albin Cassirer and
            Fan Yang and
            Kate Baumli and
            Sarah Henderson and
            Alex Novikov and
            Sergio Gómez Colmenarejo and
            Serkan Cabi and
            Caglar Gulcehre and
            Tom Le Paine and
            Andrew Cowie and
            Ziyu Wang and
            Bilal Piot and
            Nando de Freitas},
    year={2020},
    journal={arXiv preprint arXiv:2006.00979},
    url={https://arxiv.org/abs/2006.00979},
}

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

TCP/IP网络编程

TCP/IP网络编程

[韩] 尹圣雨 / 金国哲 / 人民邮电出版社 / 2014-7 / 79.00元

第一部分主要介绍网络编程基础知识。此部分主要论述Windows和Linux平台网络编程必备基础知识,未过多涉及不同操作系统特性。 第二部分和第三部分与操作系统有关。第二部分主要是Linux相关内容,而第三部分主要是Windows相关内容。从事Windows编程的朋友浏览第二部分内容后,同样可以提高技艺。 第四部分对全书内容进行总结,包含了作者在自身经验基础上总结的学习建议,还介绍了网络......一起来看看 《TCP/IP网络编程》 这本书的介绍吧!

在线进制转换器
在线进制转换器

各进制数互转换器

HTML 编码/解码
HTML 编码/解码

HTML 编码/解码

Base64 编码/解码
Base64 编码/解码

Base64 编码/解码