Skip to content

thewintersun/distributeTensorflowExample

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 

Repository files navigation

996.icu LICENSE

distributeTensorflowExample

分布式介绍中文文档

http://blog.csdn.net/luodongri/article/details/52596780

更多tensorflow和深度学习的内容,请参考我的书《tensorflow入门与实战》

这本书深度学习入门的内容占了一半,都是很基础和入门的.

如果刚入门的可以看看,自认为比网上看吴恩达的教程更容易看懂。

如果是已经比较熟悉Tensorflow和深度学习了,可以不用看了。

链接: https://item.jd.com/12307221.html

说明

这是一个最简单的分布式tensorflow的例子。

实现的功能是估计这个公式的2个参数:  Y = 2 * X + 10

要估计的参数是weight是2, biasis 是10.


程序执行的ps节点1个, worker节点2个。 执行命令示例在下面。

详细关于tensorflow的分布式示例介绍:

执行命令示例

ps 节点执行: 

CUDA_VISIBLE_DEVICES='' python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=ps --task_index=0



worker 节点执行:

CUDA_VISIBLE_DEVICES=0 python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=worker --task_index=0

CUDA_VISIBLE_DEVICES=0 python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=worker --task_index=1

Introduce

This is a most simple example for distributed tensorflow.

The task is to estimate the paramters of the formula : Y = 2 * X + 10

the paramter weight is the number 2, 

the paramter biasis is the number 10.



run example

ps server:

CUDA_VISIBLE_DEVICES='' python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=ps --task_index=0



worker server:

CUDA_VISIBLE_DEVICES=0 python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=worker --task_index=0

CUDA_VISIBLE_DEVICES=0 python distribute.py --ps_hosts=192.168.100.42:2222 --worker_hosts=192.168.100.42:2224,192.168.100.253:2225 --job_name=worker --task_index=1

About

distribute tensorflow example

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages