Interoperable and decentralized machine learning.
Branch: master
Clone or download
kiddyboots216 Status server post (#48)
Adds CloudConnectedOptimizer, which posts training statistics to status server
Latest commit 117f998 Feb 1, 2019

README.md

Decentralized Machine Learning Build Status

Instructions

To try out the project, just clone or download it and then run the following instructions on CLI:

export DB_PASS="datashark"
python setup.py develop
dml

If this is your first time running the project, it will ask you to fill out certain configurations. You can safely choose the default options by just pressing Enter multiple times, or you can manually enter your own preferences.

After this, the service will be running on your machine and you can stop it at any time by pressing Ctrl-C.

Run Decentralized Training

Prerequisites

Make sure ipfs is running in a new window with ipfs daemon

Make sure the chain is running on another window with node app_trivial.js

Creating a new DML Session

python utils_submit_dml_session_py.py

The above command will submit a small DML session to the blockchain, which your running service will be able to pick up on and train. To submit more complex sessions to the blockchain, please use the Explora frontend client.