Skip to content

An early prototype for a product intended to help you manage machine learning models end to end

License

Notifications You must be signed in to change notification settings

genixpro/electric-brain

Repository files navigation

Electric Brain

Installation

Thank you for downloading Electric Brain!

Dependencies

Electric Brain has several major dependencies:

  • NodeJS >= 4.2.6
  • MongoDB >= 2.6
  • RabbitMQ
  • TensorFlow
  • Graphviz
  • SQLLite3

Installation of these four tools will vary greatly depending on your platform.

Ubuntu 16.04

On Ubuntu 16.04, several of the required tools can be installed with their required versions very easily from the mainline package repository.

user@machine:~$ sudo apt install mongodb rabbitmq-server graphviz git build-essential libsqlite3-dev python3-pip -y

Tensorflow must be install separately, using Pip.

user@machine:~$ sudo pip3 install tensorflow

After installing Tensorflow, you will need to manually install several Electric Brain dependencies in Lua via luarocks:

user@machine:~$ sudo luarocks install json
user@machine:~$ sudo luarocks install distlearn
user@machine:~$ sudo luarocks install ipc
user@machine:~$ sudo luarocks install rnn
user@machine:~$ sudo luarocks install underscore
user@machine:~$ sudo luarocks install luasocket
user@machine:~$ sudo luarocks install lsqlite3

Now you can finally install Electric Brain via NPM. It must be installed globally using the -g command

user@machine:~$ sudo npm install electricbrain -g

Operation

The Electric Brain infrastructure is divided into two components, the API and frontend application server, and backend workers which do processing. Both of them are required for Electric Brain to function properly, and they both require connections to RabbitMQ and MongoDB. You can have as many of each running as you want.

To run the API server, simply run:

user@machine:~$ ebapi

To run the worker, simply run:

user@machine:~$ ebworker

You can also run an all-in-one application server using:

user@machine:~$ ebapplication

About

An early prototype for a product intended to help you manage machine learning models end to end

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published