在带有烧瓶的LAMP（python）堆栈上部署keras（带有theano后端）？(Deploy keras (with theano backend) on a LAMP (python) stac

在带有烧瓶的LAMP（python）堆栈上部署keras（带有theano后端）？(Deploy keras (with theano backend) on a LAMP (python) stack with flask?)

我正在尝试在计算引擎上部署带有keras和theano的Flask Web应用程序。 当我import keras并运行sudo service apache2 restart时，服务器进入循环（“等待IP地址”）。 error.log给出以下错误

[:error] [pid 1588:tid 140666780755712] Using Theano backend.

我在ubuntu 14.04

I am trying to deploy a Flask web app with keras and theano on compute engine. When I import keras and run sudo service apache2 restart, the server goes into a loop ("waiting for IP address"). The error.log gives the following error

[:error] [pid 1588:tid 140666780755712] Using Theano backend.

I am on ubuntu 14.04

最满意答案

我让keras在计算引擎上工作。 要导入繁重的库，我们需要在.conf文件中设置某些参数。

将以下代码添加到/etc/apache2/sites-enables/yourapp.conf文件中：

WSGIDaemonProcess yourapp user=you group=somegroup processes=1 threads=5 <Directory /home/directory/to/webserver/html/appdir/app.wsgi> WSGIProcessGroup yourapp WSGIApplicationGroup %{GLOBAL} Order deny,allow Allow from all </Directory>

有关详细信息，请参阅此链接 - http://cmry.github.io/notes/bottle

I got keras working on compute engine. To import heavy libraries we need to set certain parameters in .conf file.

Add the following code to your /etc/apache2/sites-enables/yourapp.conf file:

WSGIDaemonProcess yourapp user=you group=somegroup processes=1 threads=5 <Directory /home/directory/to/webserver/html/appdir/app.wsgi> WSGIProcessGroup yourapp WSGIApplicationGroup %{GLOBAL} Order deny,allow Allow from all </Directory>

Refer this link for details - http://cmry.github.io/notes/bottle